No entries match these filters. Clear filters
Findings marked unmapped when taxonomy code cannot be resolved; taxonomy snapshot overridable at runtime for custom classifications.
Labels findings whose code cannot be resolved — such as from a custom rule pack without a taxonomy selection or a code absent from a custom taxonomy — as unmapped instead of guessing a code. Not checked yet Supports runtime override of the bundled taxonomy snapshot for custom organizational classifications. Not checked yet
2 details in Everything
Skill Scanner 2.3.1 ships full AITech taxonomy definitions via get_aitech_definition() and updates stale LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. provider defaults.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.3.1 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.3.1
└──▷ USE IT
Look up the human-readable definition of an AITech code surfaced in a finding, without leaving Python.
python
copy
from skill_scanner import get_aitech_definition, get_aisubtech_definition
print(get_aitech_definition('AITech-1.1'))
print(get_aisubtech_definition('AITech-1.1.2'))
Adds get_aitech_definition() and get_aisubtech_definition() functions to look up the bundled definition for any AITech objective, technique, or sub-technique code directly from the snapshot.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet Updates the --llm-provider openrouter default model to openrouter/openai/gpt-5.6-sol and the gcp-vertex default to vertex_ai/gemini-3.8-flash to replace retired models. Not checked yet Ships a generated taxonomy reference covering all 19 objectives, 44 techniques, and 130 sub-techniques, with each entry in threats.py carrying a full description and AI Defense labels. Not checked yet
Adds get_aitech_definition() and get_aisubtech_definition() functions to look up the bundled definition for any AITech objective, technique, or sub-technique code directly from the snapshot.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet
2 more in Everything
└──▷ BREAKING ON UPGRADE
! scan_metadata.taxonomy.snapshot_sha256 changes from 3df746f0… to 413e8ce9…; anything that pins this hash must be updated.! OpenRouter OpenRouter A unified API gateway that routes requests to dozens of LLM providers under one endpoint, letting cyber tools swap or compare models without rewriting provider-specific integration code. users relying on the default model now get openrouter/openai/gpt-5.6-sol instead of openrouter/openai/gpt-4; set SKILL_SCANNER_LLM_MODEL to pin a specific model.! Vertex AI users relying on the default model now get vertex_ai/gemini-3.8-flash instead of vertex_ai/gemini-1.5-pro; set SKILL_SCANNER_LLM_MODEL to pin a specific model.
Skill Scanner 2.3.0 maps every finding to Cisco AI taxonomy codes, adds ACTIVE_SENSITIVE_EXFILTRATION detection, and surfaces SYMLINK_NOT_SCANNED and SKILL_LOAD_FAILED findin︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.3.0 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.3.0
└──▷ TRY IT
Review taxonomy metadata on a scan to see AITech/AISubtech codes and MITRE ATLAS MITRE ATLAS A MITRE-maintained knowledge base of adversarial tactics and techniques targeting AI and machine learning systems, modelled on ATT&CK, giving cyber tools a structured taxonomy for AI-specific threat modelling. crosswalks alongside each finding.
copy
$ skill-scanner scan <skill_directory> --format json | jq '.findings[] | {title, aitech, aitech_name, aisubtech, aisubtech_name, framework_mappings, taxonomy_status}'
Adds aitech, aitech_name, aisubtech, aisubtech_name, aitech_mappings, aisubtech_mappings, framework_mappings, taxonomy_status, and taxonomy_source fields to every finding, plus scan_metadata.taxonomy counts and SARIF SARIF Static Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. result properties, mapping to OWASP OWASP A nonprofit foundation that publishes open standards, guides, and ranked vulnerability lists—most famously the OWASP Top 10—giving cyber tools a vendor-neutral benchmark to test or report against. , MITRE MITRE A US federally funded research and development center that maintains public frameworks—ATT&CK, CVE, D3FEND—that cyber tools use to classify threats, vulnerabilities, and defensive techniques in a common language. ATLAS, and NIST AML AML Anti-Money Laundering: a set of laws, regulations, and procedures that financial institutions must follow to detect and report suspicious financial activity; cyber tools use it to automate transaction monitoring and compliance workflows. crosswalks from a hash-pinned taxonomy 2.0.0 snapshot bundled with the scanner.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet Adds new detection ACTIVE_SENSITIVE_EXFILTRATION to catch plain-language instructions that read a local credential file and send it to a remote host. Not checked yet Reports symlinks among a skill's files as SYMLINK_NOT_SCANNED findings (HIGH when the target leaves the skill root) and skills that fail to load as SKILL_LOAD_FAILED findings in scan-all, instead of silently skipping them. Not checked yet Extends frontmatter scanning to cover all SKILL.md values and comments — including quoted and complex keys, YAML aliases, merge keys, escape sequences, !!binary values, and multi-line folded values — except the allowed-tools value. Not checked yet
Adds aitech, aitech_name, aisubtech, aisubtech_name, aitech_mappings, aisubtech_mappings, framework_mappings, taxonomy_status, and taxonomy_source fields to every finding, plus scan_metadata.taxonomy counts and SARIF SARIF Static Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. result properties, mapping to OWASP OWASP A nonprofit foundation that publishes open standards, guides, and ranked vulnerability lists—most famously the OWASP Top 10—giving cyber tools a vendor-neutral benchmark to test or report against. , MITRE MITRE A US federally funded research and development center that maintains public frameworks—ATT&CK, CVE, D3FEND—that cyber tools use to classify threats, vulnerabilities, and defensive techniques in a common language. ATLAS, and NIST AML AML Anti-Money Laundering: a set of laws, regulations, and procedures that financial institutions must follow to detect and report suspicious financial activity; cyber tools use it to automate transaction monitoring and compliance workflows. crosswalks from a hash-pinned taxonomy 2.0.0 snapshot bundled with the scanner.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet Adds new detection ACTIVE_SENSITIVE_EXFILTRATION to catch plain-language instructions that read a local credential file and send it to a remote host. Not checked yet
2 more in Everything
└──▷ BREAKING ON UPGRADE
! Taxonomy codes follow taxonomy 2.0.0: jailbreak rules move from AITech-2.1 to AITech-1.1 / AISubtech-1.1.5, and hardcoded secrets from AISubtech-8.2.2 to AISubtech-14.1.1. Any filters on 1.x codes must be updated.! The AITech-99.9 / AISubtech-99.9.9 placeholder codes are no longer emitted.! scan-all now reports folders with unloadable skills or symlinks as findings instead of skipping them, so SARIF output, markdown output, and --fail-on-severity exit codes include them.
Skill Scanner 2.2.2 adds whole-environment dump detection and a new SKILL_SCANNER_ADJUDICATOR_LLM_MAX_TOKENS env var for LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. output control.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.2.2 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.2.2
Adds SKILL_SCANNER_ADJUDICATOR_LLM_MAX_TOKENS environment variable to set a separate output-token limit for the adjudicator LLM.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet Adds SKILL_SCANNER_ADJUDICATOR_LLM_REASONING_EFFORT environment variable to override the adjudicator's reasoning mode (disabled by default in this release). Not checked yet Adds DATA_EXFIL_ENV_DUMP detection rule for whole process-environment value dumps in Python, JavaScript/TypeScript, and shell code, with targeted reads, keys-only enumeration, literal examples, and comments kept below the HIGH rule boundary. Not checked yet
3 details in Everything
└──▷ BREAKING ON UPGRADE
! The adjudicator's reasoning is now disabled by default; set SKILL_SCANNER_ADJUDICATOR_LLM_REASONING_EFFORT to restore the previous behaviour.! A shell prompt marker by itself no longer captures a command on the next line; prompt and command must share a line.
skill-scanner adds five policy presets, consensus mode, a second-pass meta-analyzer LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. , and a GitHub Actions GitHub Actions GitHub's built-in CI/CD system that runs automated workflows triggered by repository events, letting cyber tools automate testing, building, and releasing without managing separate pipeline infrastructure. reusable workflow with measured recall/FPR benchmarks.
Provides five built-in policy presets — low-noise, quiet, balanced, plus the default and a maximum-sensitivity audit preset — selectable via policy: in the GitHub Actions workflow config. Not checked yet Adds --policy low-noise and --policy balanced and --policy quiet flags to tune the trade-off between recall and false-positive rate (measured: quiet blocks 33.1% of malicious skills at 5.5% FPR; balanced achieves ~75.5% F1). Not checked yet Surfaces LLM_ANALYSIS_FAILED at INFO severity when a skill cannot be analysed by the LLM judge, allowing CI to route unreviewed skills to human review. Not checked yet New integration Supports uses: cisco-ai-defense/skill-scanner/.github/workflows/scan-skills.yml@2.2.1 GitHub Actions reusable workflow with scanner_version, policy, use_llm, llm_model, and llm_api_key inputs. Not checked yet Supports severity_overrides and suppressions configuration to remove remaining noise without disabling rules. Not checked yet
+2 more
show less
Enables consensus mode (run the LLM analyzer multiple times and keep majority-agreed findings) as an opt-in setting. Not checked yet Supports a second-pass meta-analyzer LLM (off by default) that correlates findings across analyzers and can filter judged false positives. Not checked yet
7 details in Everything
skill-scanner adds a reusable GitHub Actions GitHub Actions GitHub's built-in CI/CD system that runs automated workflows triggered by repository events, letting cyber tools automate testing, building, and releasing without managing separate pipeline infrastructure. workflow, pre-commit pre-commit An open-source framework for managing and running git hooks, maintained by the pre-commit project, that lets cyber tools enforce checks automatically on every commit before code reaches a repository. hook, .skill_scannerrc config, and exit code 2 for configuration failures.
New integration New supported platform Adds a reusable GitHub Actions workflow at cisco-ai-defense/skill-scanner/.github/workflows/scan-skills.yml with inputs scanner_version, policy, fail_on_severity, llm_model, llm_provider, llm_base_url, aws_region, skill_path, and extra_args. Not checked yet Adds a self-hosted workflow path for role-based Bedrock or Vertex AI access using OIDC OIDC OpenID Connect, a sign-in protocol built on OAuth 2.0. An application uses identity information from a trusted provider instead of maintaining a separate password for each user. (id-token: write). Not checked yet New supported platform Adds pre-commit hook support that scans only the skills touched by a commit. Not checked yet Adds .skill_scannerrc config file at the repository root with keys skills_path, policy, use_llm, llm_model, severity_threshold, and fail_fast. Not checked yet Supports environment variables SKILL_SCANNER_LLM_MODEL and SKILL_SCANNER_LLM_PROVIDER as fallbacks in the pre-commit hook. Not checked yet
+1 more
show less
Surfaces a distinct exit code 2 ('could not run as configured') distinguishing configuration failures from finding-based failures. Not checked yet
6 details in Everything
skill-scanner: Adds --use-llm flag to enable the LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge alongside deterministic rules, raising malicious-skill recall from ~8% to ~66.7% at the balanced preset.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
Vet a third-party skill before install — highest F1, block at HIGH, pipe MEDIUM+ findings to a review queue from the JSON output.
copy
$ skill-scanner scan ./downloaded-skill --use-llm --policy balanced --fail-on-severity high --format json --output scan.json
jq -r '.findings[] | select(.severity == "MEDIUM" or .severity == "HIGH" or .severity == "CRITICAL" or .rule_id == "LLM_ANALYSIS_FAILED") | "\(.severity)\t\(.rule_id)\t\(.file_path // "-")\t\(.title)"' scan.json
Run the judge on a self-hosted Gemma model so skill content never leaves your machines — useful for air-gapped or compliance-sensitive environments.
copy
$ export SKILL_SCANNER_LLM_PROVIDER=openai-compatible
export SKILL_SCANNER_LLM_BASE_URL=http://127.0.0.1:8000/v1
export SKILL_SCANNER_LLM_MODEL=gemma-4-26b-a4b
export SKILL_SCANNER_LLM_API_KEY=unused
skill-scanner scan ./skill --use-llm --policy low-noise --fail-on-severity high
Gate skill commits in a repo — enable the LLM judge for every developer via .skill_scannerrc so the pre-commit pre-commit An open-source framework for managing and running git hooks, maintained by the pre-commit project, that lets cyber tools enforce checks automatically on every commit before code reaches a repository. hook blocks on HIGH without requiring per-developer flags.
.skill_scannerrc
copy
{
"skills_path": ".claude/skills",
"policy": "low-noise",
"use_llm": true,
"llm_model": "anthropic/claude-sonnet-5-5",
"severity_threshold": "high",
"fail_fast": true
}
Adds --policy flag accepting presets balanced, low-noise, quiet, and strict to control rule severity demotion and judge confidence capping. Not checked yet Adds --fail-on-severity flag (recommended value: high) to set the severity level that fails a scan. Not checked yet Adds --llm-decompose flag for the strict hunting preset, enabling per-component LLM decomposition of a skill. Not checked yet Adds --format sarif and --format json output formats to the scan and scan-all subcommands. Not checked yet Adds --output flag to write scan results to a named file (e.g., scan.json, results.sarif). Not checked yet
+8 more
show less
Adds --use-virustotal, --use-aidefense, and --use-osv flags to opt in to external analyzers (all off by default). Not checked yet Adds scan-all subcommand with --recursive to batch-scan a directory of skills. Not checked yet Adds environment variables SKILL_SCANNER_LLM_API_KEY, SKILL_SCANNER_LLM_MODEL, SKILL_SCANNER_LLM_PROVIDER, and SKILL_SCANNER_LLM_BASE_URL to configure the LLM provider without hardcoding credentials. Not checked yet Adds a pre-commit hook (id: skill-scanner) installable via .pre-commit-config.yaml or skill-scanner-pre-commit --install, scanning only commit-touched skills. Not checked yet Adds AWS Bedrock provider support via pip install 'cisco-ai-skill-scanner[bedrock]' and SKILL_SCANNER_LLM_MODEL values prefixed bedrock-mantle/. Not checked yet Adds REST API REST API A web API design style where clients interact with resources over HTTP using standard methods like GET, POST, and DELETE, giving cyber tools a predictable, language-agnostic way to integrate with external services. server (skill-scanner-api --port 8000) exposing a POST /scan-upload endpoint accepting a zipped skill file plus policy and use_llm form fields. Not checked yet Adds Python API via skill_scanner.SkillScanner, skill_scanner.core.analyzer_factory.build_analyzers, and skill_scanner.core.scan_policy.ScanPolicy (with ScanPolicy.from_preset() and ScanPolicy.from_yaml() ) for programmatic scanning. Not checked yet Supports local self-hosted judge via SKILL_SCANNER_LLM_PROVIDER=openai-compatible + SKILL_SCANNER_LLM_BASE_URL, keeping skill content entirely on-machine; tested with Gemma 4 26B-A4B served by vLLM. Not checked yet
13 details in Everything
skill-scanner: Adds --use-llm flag to invoke an LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge that reads a skill for intent, controlled by SKILL_SCANNER_LLM_MODEL and SKILL_SCANNER_LLM_API_KEY.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
Scan a third-party skill with the LLM judge, behavioral dataflow, and OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. dependency checks — fail CI on any HIGH finding and emit a SARIF SARIF Static Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. report for upload to your SAST SAST Static application security testing, analyzing source code or compiled code for vulnerabilities without running the application. It helps developers find problems before deployment. pipeline.
copy
$ skill-scanner scan /path/to/skill --use-llm --use-behavioral --use-osv --policy balanced --fail-on-severity high --format sarif --output results.sarif
Generate a custom low-noise policy YAML as a starting point, then tune it and point your scan at the result to minimise false positives on first-party skills.
copy
$ skill-scanner generate-policy --preset low-noise -o my_policy.yaml
skill-scanner scan /path/to/skill --use-llm --policy my_policy.yaml --fail-on-severity high
Drop skill-scanner into a GitHub Actions GitHub Actions GitHub's built-in CI/CD system that runs automated workflows triggered by repository events, letting cyber tools automate testing, building, and releasing without managing separate pipeline infrastructure. workflow to automatically scan all skills on every push, using the reusable workflow with the LLM judge and a pinned scanner version.
yaml
copy
uses: cisco-ai-defense/skill-scanner/.github/workflows/scan-skills.yml@2.2.1
with:
scanner_version: "2.2.1"
policy: low-noise
fail_on_severity: high
use_llm: true
llm_model: anthropic/claude-sonnet-5-5
secrets:
llm_api_key: ${{ secrets.SKILL_SCANNER_LLM_API_KEY }}
Adds generate-policy --preset <name> -o <file> subcommand to scaffold a custom policy YAML from any built-in preset. Not checked yet Adds --use-behavioral flag to enable Python dataflow analysis. Not checked yet Adds --use-osv flag to check skills against known-vulnerable dependencies. Not checked yet Adds --format sarif output option (alongside existing formats) for SARIF-compatible results, used with --output <file>. Not checked yet Adds --format html output option to generate a self-contained HTML report. Not checked yet
+1 more
show less
Adds an interactive wizard mode to scan that guides selection of target, analyzers, policy, and output format. Not checked yet
6 details in Everything
skill-scanner: Adds Homebrew tap for macOS installs via brew tap cisco-ai-defense/skill-scanner and brew install cisco-ai-defense/skill-scanner/skill-scanner (macOS 14+).︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
Install on macOS via Homebrew and immediately pin the LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge to a specific model for reproducible CI results.
copy
$ brew tap cisco-ai-defense/skill-scanner https://github.com/cisco-ai-defense/skill-scanner
brew install cisco-ai-defense/skill-scanner/skill-scanner
export SKILL_SCANNER_LLM_MODEL="anthropic/claude-sonnet-5-5"
Route the LLM judge through a local vLLM or OpenAI-compatible gateway instead of a cloud provider.
copy
$ export SKILL_SCANNER_LLM_PROVIDER=openai-compatible
export SKILL_SCANNER_LLM_MODEL="any name"
Install Google AI Studio / Gemini support and point the judge at a Gemini model for air-gapped or cost-sensitive environments.
copy
$ pip install "cisco-ai-skill-scanner[google]"
export SKILL_SCANNER_LLM_MODEL="gemini/<model>"
Prebuilt wheels now available for glibc Linux x86-64/ARM64 ARM64 A 64-bit CPU architecture from ARM Holdings, used in mobile chips, Apple Silicon, and cloud servers; cyber tools target it to run natively on those platforms instead of relying on emulation. , macOS x86-64/ARM64 (macOS 14+), and Windows x86-64; source distribution requires Go 1.27.1+ for the CEL CEL Common Expression Language, an open-source expression language maintained by Google, used to embed lightweight, safe policy evaluation logic into tools without running a full scripting engine. helper. Not checked yet
1 detail in Everything
api-docs-049d522bbb91
Oct 5, 2026
·
issue 095
└──▷ TRY IT
Set the LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. model and run a scan so the LLM judge can evaluate results — the scan fails fast with HTTP 400 if the key is missing.
copy
$ export SKILL_SCANNER_LLM_MODEL=anthropic/claude-sonnet-5-5
api-docs-049d522bbb91
Oct 4, 2026
·
issue 094
skill-scanner: Adds SKILL_SCANNER_LLM_MODEL environment variable to select the LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge model, e.g.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ anthropic/claude-sonnet-5-5.
└──▷ TRY IT
Run a scan with the LLM judge enabled using Claude Sonnet, selecting the strict preset to minimize false negatives in a high-stakes pipeline.
copy
$ export SKILL_SCANNER_LLM_MODEL=anthropic/claude-sonnet-5-5
Enables an LLM judge for scan analysis; if the judge cannot be built (e.g. no API key), the request fails with HTTP 400. Not checked yet
1 detail in Everything
api-docs-049d522bbb91
Oct 3, 2026
·
issue 093
skill-scanner adds LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge, OSV.dev OSV.dev An open-source vulnerability database and API maintained by Google that aggregates CVEs and ecosystem-specific advisories in a unified schema, giving cyber tools a single feed to query for package-level vulnerability data. dependency scanning, meta-analyzer, and named scan presets︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Adds an LLM judge option (recommended) powered by a configurable model such as anthropic/claude-sonnet-5-5, set via the SKILL_SCANNER_LLM_MODEL environment variable; requests fail with HTTP 400 if the judge cannot be built (e.g.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ no API key). Not checked yet Queries OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. .dev for vulnerable exactly-pinned Python and JavaScript dependencies. Not checked yet Adds a meta-analyzer option (not recommended per Recommended Settings guidance). Not checked yet Adds named scan presets — balanced, low-noise, quiet, strict, permissive — accepted as a preset name or a YAML path. Not checked yet
4 details in Everything
skill-scanner gains decomposed LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. analysis, system-one screening, consensus runs, and expanded detection surfaces for AI skill files︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
Maximize recall on a skill when model cost is acceptable — each focus area gets its own LLM pass and results are unioned.
copy
$ skill-scanner scan ./skill --use-llm --llm-decompose
Reduce false positives in CI by requiring findings to survive across three independent LLM runs before surfacing.
copy
$ skill-scanner scan ./skill --use-llm --llm-consensus-runs 3
Run a second-pass adjudication review across all findings for deeper triage of high-stakes skills.
copy
$ skill-scanner scan ./skill --use-llm --adjudicate
Adds --llm-decompose flag to run the LLM judge once per focus area (declared purpose, policy surface, security behaviors) and union the findings, raising F1 from 61.2% to 68.3% on Gemma 4 26B and from 58.7% to 81.4% on Claude Haiku 4.5 at roughly 3x the model calls.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet Adds --system-one-endpoint and --system-one-model flags to record a calibrated risk probability for triage (advisory only; cannot change a finding, severity, or verdict). Not checked yet Adds --llm-consensus-runs to run the LLM analyzer multiple times independently and retain only majority-agreed findings, using the highest severity observed across agreeing votes. Not checked yet Adds --adjudicate flag for second-pass review across all findings (off by default; measured at a cost of 16.4 points of recall). Not checked yet Detects injection and exfiltration text in SKILL.md frontmatter description and when_to_use fields, which agents load before invocation (core rules, YARA and active-directive rules; optional packs scan the body only). Not checked yet
+2 more
show less
Detects package-registry redirection across npm/yarn/pnpm, pip/uv, Go, Cargo, RubyGems, NuGet NuGet Microsoft's package manager for .NET, hosting a public registry of reusable libraries so tools can pull in dependencies without bundling or manually maintaining third-party code. , and Docker, plus unnamed script network destinations. Not checked yet Adds low-noise profile (balanced, 11 rules, mostly flags harmless skills reported at LOW) and quiet profile (19 rules reported at LOW, suited to triage queues with little review capacity). Not checked yet
7 details in Everything
skill-scanner adds OSV.dev OSV.dev An open-source vulnerability database and API maintained by Google that aggregates CVEs and ecosystem-specific advisories in a unified schema, giving cyber tools a single feed to query for package-level vulnerability data. dependency advisories, second-pass LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. validation, and five built-in scan presets︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Adds second-pass LLM validation/filtering (off by default; not recommended) to reduce false positives. Not checked yet Adds five built-in presets including default, low-noise, and quiet review-capacity packs. Not checked yet
2 details in Everything
skill-scanner FAQ reveals LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge, policy presets, OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. /VirusTotal/Cisco AI Defense integrations, and multi-cloud LLM provider support︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ USE IT
Scan third-party skills with the highest-F1 preset in CI, failing on MEDIUM or above — 66.7% recall at 15.4% false-positive rate on held-out splits.
.github/workflows/skill-scan.yml
copy
uses: cisco-ai-defense/skill-scanner/.github/workflows/scan-skills.yml@2.2.0
with:
scanner_version: "2.2.0"
policy: balanced
use_llm: true
llm_model: anthropic/claude-sonnet-5-5
secrets:
llm_api_key: ${{ secrets.SKILL_SCANNER_LLM_API_KEY }}
Use the quiet preset with the LLM judge to minimise false positives when gating production deploys — blocks 33.1% of malicious skills at a 5.5% false-positive rate at HIGH.
.github/workflows/skill-scan.yml
copy
uses: cisco-ai-defense/skill-scanner/.github/workflows/scan-skills.yml@2.2.0
with:
scanner_version: "2.2.0"
policy: quiet
use_llm: true
llm_model: anthropic/claude-sonnet-5-5
secrets:
llm_api_key: ${{ secrets.SKILL_SCANNER_LLM_API_KEY }}
Supports SKILL_SCANNER_LLM_PROVIDER=openai-compatible to connect the LLM judge to any OpenAI-compatible gateway or local server.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet Adds --policy low-noise, --policy balanced, and --policy quiet presets to tune false-positive/recall trade-offs for own vs. third-party skills. Not checked yet Integrates OSV.dev OSV.dev An open-source vulnerability database and API maintained by Google that aggregates CVEs and ecosystem-specific advisories in a unified schema, giving cyber tools a single feed to query for package-level vulnerability data. for known-vulnerable Python and JavaScript dependency-pin checks (network access required, no key needed). Not checked yet Integrates VirusTotal and Cisco AI Defense as optional cloud-based binary and content scanning backends (each requires its own API key). Not checked yet Supports Anthropic (anthropic/claude-sonnet-5-5, default), OpenAI, AWS Bedrock, Google Vertex AI and Gemini, Azure OpenAI, OpenRouter OpenRouter A unified API gateway that routes requests to dozens of LLM providers under one endpoint, letting cyber tools swap or compare models without rewriting provider-specific integration code. , OrcaRouter, Ollama, and the on-device Apple Foundation Model Experimental as LLM judge providers. Not checked yet
+4 more
show less
Adds five built-in policy presets — default, low-noise, quiet, and two additional named presets — configurable via policy in YAML configuration, controlling which rules fire, severity levels, and file limits. Not checked yet Surfaces an LLM_ANALYSIS_FAILED INFO finding when the LLM judge cannot read a skill, allowing CI to route unanalysed skills to human review. Not checked yet Raises deterministic-rules-only recall from 8% to ~67% by enabling the LLM judge; consensus mode multiplies judge runs to raise agreement threshold at proportional cost. Not checked yet Runs on Linux x86-64/ARM64 ARM64 A 64-bit CPU architecture from ARM Holdings, used in mobile chips, Apple Silicon, and cloud servers; cyber tools target it to run natively on those platforms instead of relying on emulation. and Windows x86-64 pre-built; other platforms build from source requiring Go 1.27.1+. Not checked yet
Supports SKILL_SCANNER_LLM_PROVIDER=openai-compatible to connect the LLM judge to any OpenAI-compatible gateway or local server.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet Adds --policy low-noise, --policy balanced, and --policy quiet presets to tune false-positive/recall trade-offs for own vs. third-party skills. Not checked yet Integrates OSV.dev OSV.dev An open-source vulnerability database and API maintained by Google that aggregates CVEs and ecosystem-specific advisories in a unified schema, giving cyber tools a single feed to query for package-level vulnerability data. for known-vulnerable Python and JavaScript dependency-pin checks (network access required, no key needed). Not checked yet Integrates VirusTotal and Cisco AI Defense as optional cloud-based binary and content scanning backends (each requires its own API key). Not checked yet Supports Anthropic (anthropic/claude-sonnet-5-5, default), OpenAI, AWS Bedrock, Google Vertex AI and Gemini, Azure OpenAI, OpenRouter OpenRouter A unified API gateway that routes requests to dozens of LLM providers under one endpoint, letting cyber tools swap or compare models without rewriting provider-specific integration code. , OrcaRouter, Ollama, and the on-device Apple Foundation Model Experimental as LLM judge providers. Not checked yet
+3 more
show less
Adds five built-in policy presets — default, low-noise, quiet, and two additional named presets — configurable via policy in YAML configuration, controlling which rules fire, severity levels, and file limits. Not checked yet Surfaces an LLM_ANALYSIS_FAILED INFO finding when the LLM judge cannot read a skill, allowing CI to route unanalysed skills to human review. Not checked yet Runs on Linux x86-64/ARM64 ARM64 A 64-bit CPU architecture from ARM Holdings, used in mobile chips, Apple Silicon, and cloud servers; cyber tools target it to run natively on those platforms instead of relying on emulation. and Windows x86-64 pre-built; other platforms build from source requiring Go 1.27.1+. Not checked yet
1 more in Everything
skill-scanner gains documented tuning workflow: severity presets, scoped suppressions, judge confidence caps, and JSON/SARIF SARIF Static Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. output fields︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
Gate a CI pipeline at HIGH severity and fail explicitly if the judge could not analyse any skill.
copy
$ skill-scanner scan-all ./skills --recursive --use-llm --policy low-noise --format json --output results.json
jq -e '[.results[].findings[] | select(.rule_id == "LLM_ANALYSIS_FAILED")] | length == 0' results.json \
|| { echo "some skills were not analysed by the judge: review them"; exit 1; }
Start from the low-noise preset and cap the judge's contextual findings plus re-rate a noisy rule, so false positives drop without silencing detections elsewhere.
copy
$ skill-scanner generate-policy --preset low-noise -o my-policy.yaml
Apply severity override, scoped suppression, and judge confidence caps in the generated policy file before scanning.
my-policy.yaml
copy
severity_overrides:
- rule_id: FILE_MAGIC_MISMATCH
severity: LOW
reason: "mostly a text label in Markdown on our skills"
suppressions:
- rule_id: HOMOGLYPH_ATTACK
skills: ["docs-translator"]
reason: "Skill legitimately contains Cyrillic prose"
expires: 2030-12-31
llm_analysis:
low_confidence_max_severity: LOW
contextual_risk_max_severity: LOW
max_total_prompt_chars: 200000
max_code_file_chars: 50000
Adds --fail-on-severity flag to gate CI jobs at a chosen severity level (e.g., --fail-on-severity high).︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet Adds --format json output with findings[], suppressed_findings[], is_safe, and max_severity fields per skill; scan-all wraps results in results[] under a summary with counts by severity. Not checked yet Adds LLM_ANALYSIS_FAILED and LLM_CONTEXT_BUDGET_EXCEEDED INFO-severity findings to surface skills the judge could not fully read, allowing CI pipelines to fail explicitly on unanalysed skills. Not checked yet Adds severity_overrides policy key to re-rate noisy rules to LOW without disabling them globally, keeping detections intact elsewhere. Not checked yet Adds suppressions policy key supporting scoped suppression by skills, paths, rule_id, reason, and expires; suppressed findings appear in suppressed_findings in JSON and as dismissed in SARIF. Not checked yet
+7 more
show less
Adds llm_analysis.low_confidence_max_severity and llm_analysis.contextual_risk_max_severity policy keys to cap the judge's weakest findings at LOW. Not checked yet Adds llm_analysis.max_total_prompt_chars, max_instruction_body_chars, max_code_file_chars, and max_referenced_file_chars policy keys to control the judge's prompt budget. Not checked yet Adds skill-scanner generate-policy --preset <name> -o <file> subcommand to bootstrap a policy file from a named preset (low-noise, balanced, or quiet). Not checked yet Adds skill-scanner configure-policy -i <file> -o <file> interactive subcommand for editing policy files in place. Not checked yet Adds disabled_rules policy key to switch a rule off globally (noted as a last resort vs. scoped suppression). Not checked yet Documents exit code semantics: 0 = nothing reached threshold, 1 = finding at or above --fail-on-severity, 2 = policy/configuration error (e.g., missing API key); same configuration error returns HTTP 400 over the REST API REST API A web API design style where clients interact with resources over HTTP using standard methods like GET, POST, and DELETE, giving cyber tools a predictable, language-agnostic way to integrate with external services. . Not checked yet SARIF output reports suppressed findings as dismissed, preserving the record in Code Scanning. Not checked yet
12 details in Everything
skill-scanner: Enables an LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge via SKILL_SCANNER_LLM_MODEL (e.g.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ anthropic/claude-sonnet-5-5) to perform AI-assisted evaluation; requests fail with HTTP 400 if the judge cannot be built
└──▷ TRY IT
Enable the LLM judge with a specific model before running a scan, so AI-assisted evaluation is applied to results
copy
$ export SKILL_SCANNER_LLM_MODEL=anthropic/claude-sonnet-5-5
skill-scanner Python SDK gains named scan presets (balanced, low-noise, quiet, strict, permissive) and LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. -backed analysis via SKILL_SCANNER_LLM_* env vars.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Adds use_llm=True analyzer option to enable LLM-backed analysis, pulling model and API key from SKILL_SCANNER_LLM_* environment variables. Not checked yet
1 detail in Everything
skill-scanner: Adds .skill_scannerrc config file (repository root) with keys skills_path, policy, use_llm, llm_model, severity_threshold, and fail_fast to configure the LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge and scan behaviour.
└──▷ TRY IT
Gate PRs on HIGH-severity findings in your own skills, surfacing MEDIUM findings as inline annotations for reviewers.
copy
$ # .github/workflows/skill-scan.yml
jobs:
skill-scan:
uses: cisco-ai-defense/skill-scanner/.github/workflows/scan-skills.yml@2.2.0
secrets:
SKILL_SCANNER_LLM_API_KEY: ${{ secrets.SKILL_SCANNER_LLM_API_KEY }}
with:
scanner_version: "2.2.0"
policy: low-noise
fail_on_severity: high
llm_model: anthropic/claude-sonnet-5-5
permissions:
actions: read
security-events: write
contents: read
Configure the LLM judge and severity gate for local and pre-commit pre-commit An open-source framework for managing and running git hooks, maintained by the pre-commit project, that lets cyber tools enforce checks automatically on every commit before code reaches a repository. runs without repeating flags on every invocation.
.skill_scannerrc
copy
{
"skills_path": ".claude/skills",
"policy": "low-noise",
"use_llm": true,
"llm_model": "anthropic/claude-sonnet-5-5",
"severity_threshold": "high",
"fail_fast": true
}
Adds pre-commit hook integration (pinned via rev: 2.2.0) that scans only the skills touched by a commit, blocking the commit with exit code 2 if the judge cannot be initialised. Not checked yet Adds --use-virustotal and --use-osv flags to skill-scanner scan-all for Tier 3 behavioral and reputation checks on third-party skills. Not checked yet Supports bedrock/ models in the reusable workflow via a Bedrock API key (SKILL_SCANNER_LLM_API_KEY secret) plus aws_region input (default us-east-1), and via an OIDC OIDC OpenID Connect, a sign-in protocol built on OAuth 2.0. An application uses identity information from a trusted provider instead of maintaining a separate password for each user. role-assumed self-hosted workflow for role-based Bedrock access. Not checked yet Adds SKILL_SCANNER_LLM_MODEL and SKILL_SCANNER_LLM_PROVIDER environment variables as fallbacks for llm_model and llm_provider in pre-commit and self-hosted setups. Not checked yet Exposes --llm-decompose and LLM-limit tuning options via the allowlist in the reusable workflow's extra_args input. Not checked yet
+1 more
show less
Supports --format sarif --output results.sarif output with subsequent upload via github/codeql-action/upload-sarif@v4 in self-hosted workflows. Not checked yet
Adds pre-commit hook integration (pinned via rev: 2.2.0) that scans only the skills touched by a commit, blocking the commit with exit code 2 if the judge cannot be initialised. Not checked yet Adds --use-virustotal and --use-osv flags to skill-scanner scan-all for Tier 3 behavioral and reputation checks on third-party skills. Not checked yet Supports bedrock/ models in the reusable workflow via a Bedrock API key (SKILL_SCANNER_LLM_API_KEY secret) plus aws_region input (default us-east-1), and via an OIDC OIDC OpenID Connect, a sign-in protocol built on OAuth 2.0. An application uses identity information from a trusted provider instead of maintaining a separate password for each user. role-assumed self-hosted workflow for role-based Bedrock access. Not checked yet Adds SKILL_SCANNER_LLM_MODEL and SKILL_SCANNER_LLM_PROVIDER environment variables as fallbacks for llm_model and llm_provider in pre-commit and self-hosted setups. Not checked yet Supports --format sarif --output results.sarif output with subsequent upload via github/codeql-action/upload-sarif@v4 in self-hosted workflows. Not checked yet
1 more in Everything
skill-scanner adds OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. dependency advisory queries, System One screening, decomposed LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. passes, and new policy presets︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
Vet a third-party skill in CI/CD, failing the pipeline on any HIGH severity finding and emitting a SARIF SARIF Static Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. report for upload.
copy
$ skill-scanner scan ./my-skill --use-llm --policy low-noise --fail-on-severity high --format sarif --output results.sarif
Run a full dependency advisory check against OSV.dev OSV.dev An open-source vulnerability database and API maintained by Google that aggregates CVEs and ecosystem-specific advisories in a unified schema, giving cyber tools a single feed to query for package-level vulnerability data. alongside LLM and behavioral analysis for a thorough audit with no API key.
copy
$ skill-scanner scan ./my-skill --use-behavioral --use-llm --use-osv --use-aidefense
Generate a reusable low-noise policy file to standardise scan settings across pipelines.
copy
$ skill-scanner generate-policy -o my_policy.yaml --preset low-noise
Adds --system-one-endpoint URL flag to route scans through an optional System One screening endpoint (POST /v1/systemone); advisory only and cannot alter findings, severity, or verdict.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet Adds --system-one-model NAME flag to specify the model for the System One endpoint; required when --system-one-endpoint is set and rejected without it. Not checked yet Adds --use-aidefense flag to enable the meta-analyzer (off by default; measured at a 16.4-point recall cost). Not checked yet
3 details in Everything
skill-scanner: Adds SKILL_SCANNER_LLM_FORCE_JSON_OBJECT=true to fall back to plain JSON mode JSON mode Forcing a model's response to match a given format or JSON schema, so software can parse it reliably instead of hoping the text is well formed. for gateways that reject json_schema structured output structured output Forcing a model's response to match a given format or JSON schema, so software can parse it reliably instead of hoping the text is well formed. .
└──▷ TRY IT
Keep skill content fully on-prem by routing the LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge through a local vLLM server — no skill data leaves your machines.
copy
$ export SKILL_SCANNER_LLM_PROVIDER="openai-compatible"
export SKILL_SCANNER_LLM_BASE_URL="http://127.0.0.1:8000/v1"
export SKILL_SCANNER_LLM_MODEL="gemma-4-26b-a4b"
export SKILL_SCANNER_LLM_API_KEY="unused"
skill-scanner scan ./skill --use-llm
Scan with AWS Bedrock as the judge backend using an IAM IAM Identity and Access Management — a framework of policies and services for controlling who can authenticate to systems and what resources they are authorized to use, letting cyber tools enforce least-privilege and audit access. profile — no API key needed, and prompt budget raised for large skills.
copy
$ pip install "cisco-ai-skill-scanner[bedrock]"
export AWS_REGION="us-east-1"
export AWS_PROFILE="my-bedrock-profile"
export SKILL_SCANNER_LLM_MODEL="bedrock/us.anthropic.claude-sonnet-5-5"
skill-scanner scan ./skill --use-llm
Cap noisy low-confidence LLM findings and raise the prompt budget for large skills in your policy file.
yaml
copy
llm_analysis:
low_confidence_max_severity: LOW
contextual_risk_max_severity: LOW
max_total_prompt_chars: 200000
max_instruction_body_chars: 80000
max_code_file_chars: 40000
max_referenced_file_chars: 40000
Adds SKILL_SCANNER_LLM_USER to set the Chat Completions user field for gateways that require it for routing or billing. Not checked yet Adds SKILL_SCANNER_LLM_TEMPERATURE (accepts a number or none to always omit the parameter) to override the default temperature-0 behaviour. Not checked yet Adds SKILL_SCANNER_LLM_REPAIR_INCONSISTENT_VERDICT — set to 0 to disable verdict repair (the default escalates a self-contradicting SAFE verdict to SUSPICIOUS). Not checked yet Adds SKILL_SCANNER_LLM_MAX_TOKENS / --llm-max-tokens N (default 8192) to control output token budget for the judge. Not checked yet Adds SKILL_SCANNER_LLM_REASONING_EFFORT / --llm-reasoning-effort with levels disabled, minimal, low, medium, high, xhigh, max to control reasoning depth. Not checked yet
+7 more
show less
Adds SKILL_SCANNER_META_LLM_* variable family to override the model used by the meta-analyzer independently of the primary judge. Not checked yet Adds SKILL_SCANNER_ADJUDICATOR_LLM_MODEL to override the model used for --adjudicate independently of the primary judge. Not checked yet Supports AWS Bedrock as an LLM judge backend via pip install 'cisco-ai-skill-scanner[bedrock]', using bedrock/<model> or bedrock-mantle/<model> prefixes with standard IAM credentials. Not checked yet Supports Google Vertex AI (pip install 'cisco-ai-skill-scanner[vertex]', model prefix vertex_ai/<model>) and Google AI Studio / Gemini API (pip install 'cisco-ai-skill-scanner[google]', prefix gemini/<model>) as judge backends. Not checked yet Supports Azure OpenAI as a judge backend via pip install 'cisco-ai-skill-scanner[azure]', configured with SKILL_SCANNER_LLM_MODEL=azure/<deployment-name>, SKILL_SCANNER_LLM_BASE_URL, and SKILL_SCANNER_LLM_API_VERSION. Not checked yet Supports Ollama as a local judge backend via SKILL_SCANNER_LLM_MODEL=ollama/<model> (defaults SKILL_SCANNER_LLM_BASE_URL to http://127.0.0.1:11434); the route is loopback-only, preventing skill content from leaving the host. Not checked yet Adds experimental on-device Apple Foundation Model judge backend (SKILL_SCANNER_LLM_MODEL=apple-fm/system) for macOS 26+ with Apple Intelligence — no API key required; requires pip install 'apple-fm-sdk>=0.2.1,<0.3'. Not checked yet
12 details in Everything
skill-scanner: Adds llm_model policy YAML key to configure which LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. the judge uses (e.g. anthropic/claude-sonnet-5-5)
└──▷ TRY IT
Gate CI on HIGH findings while using the LLM judge across a skills directory, with policy defined in a shared YAML file
copy
$ skill-scanner scan-all ./skills --use-llm --policy .github/scan-policy.yaml --fail-on-severity high
Generate an editable policy from the low-noise preset, then customise it to suppress a noisy rule for one skill with an expiry date
copy
$ skill-scanner generate-policy --preset low-noise -o .github/scan-policy.yaml
Tune a policy YAML to cap weak LLM findings, suppress a false-positive for a specific skill, and pin the judge model
.github/scan-policy.yaml
copy
use_llm: true
llm_model: anthropic/claude-sonnet-5-5
low_confidence_max_severity: LOW
contextual_risk_max_severity: LOW
suppressions:
- rule_id: HOMOGLYPH_ATTACK
skills: ["docs-translator"]
reason: "Skill legitimately contains Cyrillic prose"
expires: 2030-12-31
Adds --fail-on-severity flag to skill-scanner scan-all, enabling severity-gated CI — e.g. --fail-on-severity high to block on HIGH findings Not checked yet
1 detail in Everything
skill-scanner: Adds --use-llm flag to enable the LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge alongside deterministic rules, raising recall of malicious skills from ~8% to 66.7% at MEDIUM+ threshold.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
Vet a third-party skill before install, blocking on HIGH and writing MEDIUM+ findings to a JSON review queue.
copy
$ skill-scanner scan ./downloaded-skill --use-llm --policy balanced --fail-on-severity high --format json --output scan.json
jq -r '.findings[] | select(.severity == "MEDIUM" or .severity == "HIGH" or .severity == "CRITICAL" or .rule_id == "LLM_ANALYSIS_FAILED") | "\(.severity)\t\(.rule_id)\t\(.file_path // "-")\t\(.title)"' scan.json
Run the LLM judge on a self-hosted vLLM model so skill content never leaves your machines.
copy
$ export SKILL_SCANNER_LLM_PROVIDER=openai-compatible
export SKILL_SCANNER_LLM_BASE_URL=http://127.0.0.1:8000/v1
export SKILL_SCANNER_LLM_MODEL=gemma-4-26b-a4b
export SKILL_SCANNER_LLM_API_KEY=unused
skill-scanner scan ./skill --use-llm --policy low-noise --fail-on-severity high
Wire skill-scanner into a repo so every commit only scans touched skills, with the LLM judge enabled via a repo-root config file.
.skill_scannerrc
copy
{
"skills_path": ".claude/skills",
"policy": "low-noise",
"use_llm": true,
"llm_model": "anthropic/claude-sonnet-5-5",
"severity_threshold": "high",
"fail_fast": true
}
Adds --llm-decompose flag for deep hunting mode alongside the strict preset and all rule packs. Not checked yet Adds --format flag supporting json and sarif output formats, and --output flag to write results to a file. Not checked yet Adds --use-virustotal, --use-aidefense, and --use-osv flags to opt in to external analyzers (off by default to keep skill content on-machine). Not checked yet Adds --recursive flag to scan-all for scanning an entire registry or directory tree. Not checked yet Adds a REST API REST API A web API design style where clients interact with resources over HTTP using standard methods like GET, POST, and DELETE, giving cyber tools a predictable, language-agnostic way to integrate with external services. server via skill-scanner-api --port 8000 exposing a POST /scan-upload endpoint that accepts a skill ZIP, policy, and use_llm parameters. Not checked yet
+3 more
show less
Adds a Python SDK with SkillScanner, build_analyzers, and ScanPolicy (from skill_scanner, skill_scanner.core.analyzer_factory, skill_scanner.core.scan_policy) for programmatic scanning. Not checked yet The low-noise preset demotes the 11 rules most likely to flag a harmless skill and caps low-confidence judge findings to LOW severity, reducing the review queue by ~3 percentage points versus balanced. Not checked yet The quiet preset demotes 19 rules and additionally caps CONTEXTUAL_RISK judge findings to LOW, achieving 7.2% FPR while keeping demoted findings visible at LOW. Not checked yet
8 details in Everything
skill-scanner: Adds --use-behavioral flag to enable Python dataflow analysis during a scan.
└──▷ TRY IT
Scan a third-party skill with the LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge, all analyzers enabled, and fail CI on any HIGH finding — outputting a SARIF SARIF Static Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. file for ingestion into a SIEM SIEM Security information and event management, collecting logs from many systems and correlating events to find suspicious activity. Analysts use it to investigate alerts and incidents. or GitHub Code Scanning.
copy
$ skill-scanner scan-all ./skills --recursive --use-llm --use-behavioral --use-osv --policy balanced --fail-on-severity high --format sarif --output results.sarif
Start from a built-in preset to create a tuned policy with a low false-positive rate, then scan your own skills against it.
copy
$ skill-scanner generate-policy --preset low-noise -o my_policy.yaml
skill-scanner scan /path/to/skill --use-llm --policy my_policy.yaml --fail-on-severity high
Wire skill-scanner into a GitHub Actions GitHub Actions GitHub's built-in CI/CD system that runs automated workflows triggered by repository events, letting cyber tools automate testing, building, and releasing without managing separate pipeline infrastructure. pipeline using the reusable workflow, keeping the LLM API key in repository secrets.
.github/workflows/scan-skills.yml
copy
uses: cisco-ai-defense/skill-scanner/.github/workflows/scan-skills.yml@2.2.0
with:
scanner_version: '2.2.0'
policy: low-noise
fail_on_severity: high
use_llm: true
llm_model: anthropic/claude-sonnet-5-5
secrets:
llm_api_key: ${{ secrets.SKILL_SCANNER_LLM_API_KEY }}
Adds --use-osv flag to check skills against known-vulnerable dependencies. Not checked yet Adds an interactive wizard mode to scan that guides selection of target, analyzers, policy, and output format. Not checked yet
2 details in Everything
skill-scanner: Adds Homebrew tap install via brew tap cisco-ai-defense/skill-scanner and brew install cisco-ai-defense/skill-scanner/skill-scanner for macOS 14+.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
Install skill-scanner on macOS via Homebrew for a dependency-isolated, project-independent setup.
copy
$ brew tap cisco-ai-defense/skill-scanner https://github.com/cisco-ai-defense/skill-scanner
brew install cisco-ai-defense/skill-scanner/skill-scanner
Prebuilt wheels now cover glibc Linux x86-64 and ARM64 ARM64 A 64-bit CPU architecture from ARM Holdings, used in mobile chips, Apple Silicon, and cloud servers; cyber tools target it to run natively on those platforms instead of relying on emulation. , macOS x86-64 and ARM64 (macOS 14+), and Windows x86-64; source distribution requires Go 1.27.1+ for the CEL CEL Common Expression Language, an open-source expression language maintained by Google, used to embed lightweight, safe policy evaluation logic into tools without running a full scripting engine. helper. Not checked yet
1 detail in Everything
Docs describe --use-llm flag, GitHub Actions GitHub Actions GitHub's built-in CI/CD system that runs automated workflows triggered by repository events, letting cyber tools automate testing, building, and releasing without managing separate pipeline infrastructure. /pre-commit pre-commit An open-source framework for managing and running git hooks, maintained by the pre-commit project, that lets cyber tools enforce checks automatically on every commit before code reaches a repository. CI integration, REST API REST API A web API design style where clients interact with resources over HTTP using standard methods like GET, POST, and DELETE, giving cyber tools a predictable, language-agnostic way to integrate with external services. , preset configs, and custom policy YAML for the skill-scanner CLI.
Adds --use-llm flag to enable the LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge analyzer on top of the automatically-run static, bytecode, and pipeline analyzers. Not checked yet Supports scanning a local skill, a directory of skills, or a GitHub repository as scan targets. Not checked yet Provides GitHub Actions reusable workflow and pre-commit hook for CI/CD integration. Not checked yet Exposes a REST API for reading findings and managing the review queue. Not checked yet Supports preset configs and custom policy YAML for tuning false-positive rates, with measured recall, FPR, and F1 per setup documented in Recommended Settings. Not checked yet
+1 more
show less
Supports scoped suppressions alongside presets and the LLM judge for noise reduction. Not checked yet
6 details in Everything
Docs describe --use-llm flag, GitHub Actions GitHub Actions GitHub's built-in CI/CD system that runs automated workflows triggered by repository events, letting cyber tools automate testing, building, and releasing without managing separate pipeline infrastructure. /pre-commit pre-commit An open-source framework for managing and running git hooks, maintained by the pre-commit project, that lets cyber tools enforce checks automatically on every commit before code reaches a repository. CI integration, a REST API REST API A web API design style where clients interact with resources over HTTP using standard methods like GET, POST, and DELETE, giving cyber tools a predictable, language-agnostic way to integrate with external services. , a policy YAML system, and preset configurations with measured recall/FPR/F1 rates.
Adds --use-llm flag to invoke an LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. judge that reads a skill for intent, layered on top of the automatic static, bytecode, and pipeline analyzers. Not checked yet Supports scanning a local skill, a directory of skills, or a GitHub repository as input targets. Not checked yet Provides GitHub Actions reusable workflow and a pre-commit hook for CI/CD integration. Not checked yet Exposes a REST API for reading findings and building a review queue. Not checked yet Offers presets and a custom policy YAML system with scoped suppressions to reduce false positives, with measured recall, FPR, and F1 rates published per setup. Not checked yet
5 details in Everything
Skill Scanner 2.2.0 adds low-noise/quiet presets, scoped suppressions, Apple on-device LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. , OpenAI-compatible gateway support, and two new detection rules.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.2.0 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.2.0
└──▷ USE IT
Suppress a known false positive for a specific skill and path without disabling the rule globally — ideal for reviewed test fixtures in CI.
yaml
copy
suppressions:
- rule_id: ARCHIVE_FILE_DETECTED
skills: ["my-skill"]
paths: ["**/fixtures/**/*.zip"]
severity: LOW
reason: "Reviewed test fixtures"
Enable the LLM judge in pre-commit pre-commit An open-source framework for managing and running git hooks, maintained by the pre-commit project, that lets cyber tools enforce checks automatically on every commit before code reaches a repository. so every commit to a skills repo is semantically analyzed before it lands.
.skill_scannerrc
copy
use_llm: true
llm_model: claude-sonnet-5-5
llm_provider: anthropic
Route LLM analysis through an internal OpenAI-compatible gateway instead of a first-party provider — useful in air-gapped or proxy environments.
copy
$ export SKILL_SCANNER_LLM_PROVIDER=openai-compatible
export SKILL_SCANNER_LLM_API_KEY="your_gateway_key"
skill-scanner ./my-skill --use-llm
Adds low-noise and quiet scan presets that demote high-false-positive rules and cap judge findings via low_confidence_max_severity and contextual_risk_max_severity.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet Adds scoped rule suppressions in policy YAML: target a specific skill with skills or a file path pattern with paths, so a single false positive no longer disables a rule scan-wide. Not checked yet Adds SKILL_SCANNER_LLM_PROVIDER=openai-compatible environment variable to route LLM analysis through any OpenAI-compatible gateway. Not checked yet Adds --llm-decompose flag to run one focused LLM analysis pass per concern instead of a combined pass. Not checked yet Adds use_llm, llm_model, and llm_provider keys to .skill_scannerrc so the pre-commit hook can invoke the LLM judge. Not checked yet
+6 more
show less
Adds scanner_version, llm_provider, llm_base_url, and aws_region inputs to the reusable GitHub Actions GitHub Actions GitHub's built-in CI/CD system that runs automated workflows triggered by repository events, letting cyber tools automate testing, building, and releasing without managing separate pipeline infrastructure. workflow. Not checked yet Adds new SUPPLY_CHAIN_REGISTRY_REDIRECT (HIGH) and UNDECLARED_NETWORK_DESTINATION (LOW) static detection rules. Not checked yet Adds experimental on-device Apple Foundation Model support as an LLM provider. Not checked yet Extends OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. dependency scanning to read package.json (npm) in addition to previously supported manifests. Not checked yet Extends frontmatter scanning to cover description and when_to_use fields across core rules, YARA, and active-directive rules. Not checked yet Updates default LLM model to claude-sonnet-5-5 (or bedrock/us.anthropic.claude-sonnet-5-5 on Bedrock); adds a recommended-settings guide with measured recall/FPR/F1 figures across all presets. Not checked yet
Adds low-noise and quiet scan presets that demote high-false-positive rules and cap judge findings via low_confidence_max_severity and contextual_risk_max_severity.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet Adds scoped rule suppressions in policy YAML: target a specific skill with skills or a file path pattern with paths, so a single false positive no longer disables a rule scan-wide. Not checked yet Adds SKILL_SCANNER_LLM_PROVIDER=openai-compatible environment variable to route LLM analysis through any OpenAI-compatible gateway. Not checked yet Adds --llm-decompose flag to run one focused LLM analysis pass per concern instead of a combined pass. Not checked yet Adds new SUPPLY_CHAIN_REGISTRY_REDIRECT (HIGH) and UNDECLARED_NETWORK_DESTINATION (LOW) static detection rules. Not checked yet
+1 more
show less
Adds experimental on-device Apple Foundation Model support as an LLM provider. Not checked yet
5 more in Everything
└──▷ BREAKING ON UPGRADE
! --use-llm, --use-behavioral, and --enable-meta now fail closed: if the requested analyzer cannot be built (missing API key, missing extra, unknown provider), the scan exits with code 2 and the REST API REST API A web API design style where clients interact with resources over HTTP using standard methods like GET, POST, and DELETE, giving cyber tools a predictable, language-agnostic way to integrate with external services. returns 400. Previously the scan ran rules only and exited 0.! Exit code 2 now means a usage or configuration error (e.g. unknown policy); previously this condition returned exit code 1. Exit code 1 still means findings at or above --fail-on-severity.! The default LLM model is now claude-sonnet-5-5; the previous default claude-3-5-sonnet-20241022 is retired.! Verdict repair is on by default: a SAFE verdict that lists findings is promoted to SUSPICIOUS. To restore the prior behavior, set SKILL_SCANNER_LLM_REPAIR_INCONSISTENT_VERDICT=0.! Unpinned dependencies are now reported at LOW severity instead of MEDIUM.! The interactive wizard no longer enables the meta-analyzer by default.
skill-scanner adds CEL CEL Common Expression Language, an open-source expression language maintained by Google, used to embed lightweight, safe policy evaluation logic into tools without running a full scripting engine. correlation gating and ATR rule packs with per-invocation mode override via new CLI flags
└──▷ TRY IT
Promote CEL gates to enforce mode for a single CI run to validate coverage before rolling it out globally.
copy
$ skill-scanner scan --cel-mode enforce <target>
Include the ATR rule pack alongside core rules to extend detection coverage in an on-demand scan.
copy
$ skill-scanner scan --rule-packs atr <target>
Adds --cel-mode off|shadow|enforce flag to override the active CEL gate policy for a single invocation without changing the global configuration. Not checked yet Adds --rule-packs atr flag to opt in to the ATR rule pack as an optional input outside the core + CEL release gate. Not checked yet Adds cel.mode configuration key supporting off, shadow, and enforce values to control CEL decision behavior globally. Not checked yet Adds analyzers.correlation configuration key enabling bounded source/sink and staged-behavior correlation analysis. Not checked yet
4 details in Everything
skill-scanner gains --trusted-rule-pack and --cel-mode flags for administrator-trusted schema-v2 pack loading and CEL CEL Common Expression Language, an open-source expression language maintained by Google, used to embed lightweight, safe policy evaluation logic into tools without running a full scripting engine. policy control.
└──▷ TRY IT
Validate an administrator-approved YARA/CEL pack before deploying it to production, checking for schema-v2 compliance and CEL type errors.
copy
$ skill-scanner validate-rules --trusted-rule-pack /path/to/org-pack
Run in shadow mode to observe CEL policy matches without enforcement — useful when rolling out a new rule pack to catch false positives first.
copy
$ skill-scanner validate-rules --trusted-rule-pack /path/to/org-pack --cel-mode shadow
Adds --trusted-rule-pack PATH flag to load an administrator-trusted schema-v2 signature/YARA/CEL pack; repeatable for multiple packs. Not checked yet Adds --cel-mode MODE flag to override policy CEL mode with off, shadow, or enforce. Not checked yet Extends validate-rules to validate bundled rules and optional trusted schema-v2 packs, including CEL parse/type checks. Not checked yet
Adds --trusted-rule-pack PATH flag to load an administrator-trusted schema-v2 signature/YARA/CEL pack; repeatable for multiple packs. Not checked yet Adds --cel-mode MODE flag to override policy CEL mode with off, shadow, or enforce. Not checked yet
1 more in Everything
skill-scanner adds CEL CEL Common Expression Language, an open-source expression language maintained by Google, used to embed lightweight, safe policy evaluation logic into tools without running a full scripting engine. -based correlation mode and a validate-rules subcommand with trusted rule pack support
└──▷ TRY IT
Run CEL decisions in shadow mode first to observe results without blocking scans before committing to enforce
copy
$ skill-scanner validate-rules --trusted-rule-pack /path/to/org-pack
Adds skill-scanner validate-rules --trusted-rule-pack <path> subcommand to validate rules against an org-supplied rule pack Not checked yet
1 detail in Everything
Skill Scanner 2.1.0 adds a CEL CEL Common Expression Language, an open-source expression language maintained by Google, used to embed lightweight, safe policy evaluation logic into tools without running a full scripting engine. decision layer with off/shadow/enforce modes and expands Python static detection for shell-enabled subprocess calls.
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.1.0 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.1.0
Adds CEL policy modes off, shadow, and enforce — a bounded CEL decision layer (backed by cel-go v0.32.0) that correlates typed analyzer facts after deterministic detection and before optional LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. /Meta analysis. Not checked yet Adds a versioned protobuf fact contract (skill_scanner.semantic.v1.ScanFacts) with strict schema-v2 pack validation, immutable compiled rule generations, expression hashes, and fail-open telemetry. Not checked yet Expands static detection to identify shell-enabled subprocess calls through aliases and class scopes, constructed sensitive-file reads, import aliases, and commands hidden by supported repeating-XOR encoding. Not checked yet Adds pipeline detection for recovered runtime commands. Not checked yet
Adds CEL policy modes off, shadow, and enforce — a bounded CEL decision layer (backed by cel-go v0.32.0) that correlates typed analyzer facts after deterministic detection and before optional LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. /Meta analysis. Not checked yet Adds a versioned protobuf fact contract (skill_scanner.semantic.v1.ScanFacts) with strict schema-v2 pack validation, immutable compiled rule generations, expression hashes, and fail-open telemetry. Not checked yet Adds pipeline detection for recovered runtime commands. Not checked yet
1 more in Everything
skill-scanner 2.0.14 adds semantic data exfiltration data exfiltration Moving data out of a system or organization without authorization. It can happen through network transfers, removable media or other channels. detection
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.14 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.14
skill-scanner adds pre-commit pre-commit An open-source framework for managing and running git hooks, maintained by the pre-commit project, that lets cyber tools enforce checks automatically on every commit before code reaches a repository. hook, incremental CI scans, and an expanded CLI allowlist for GitHub Actions GitHub Actions GitHub's built-in CI/CD system that runs automated workflows triggered by repository events, letting cyber tools automate testing, building, and releasing without managing separate pipeline infrastructure. workflows
└──▷ TRY IT
Install the pre-commit hook so commits are blocked when changed skills exceed the configured severity threshold.
copy
$ skill-scanner-pre-commit --install
Run incremental skill scans in CI between a base and head commit, limiting work to only the changed files in a PR.
copy
$ pre-commit run skill-scanner --from-ref "$BASE_SHA" --to-ref "$HEAD_SHA"
Adds skill-scanner-pre-commit --install to install a pre-commit hook that maps each changed path to its nearest parent SKILL.md, scans each affected skill once, and blocks commits at the configured severity. Not checked yet Adds --scan-all flag to the pre-commit hook to scan every configured skill instead of only those affected by changed paths. Not checked yet Supports incremental CI scans via pre-commit run skill-scanner --from-ref "$BASE_SHA" --to-ref "$HEAD_SHA" to scan only changed files between two revisions. Not checked yet Expands the reusable workflow's CLI allowlist to include --rule-packs, --use-aidefense, --use-trigger, and --enable-meta, plus LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. limits and repository-local policy/taxonomy paths; unknown flags are rejected before the scan runs. Not checked yet Documents that --use-osv and --adjudicate are not accepted by the reusable workflow and require the self-hosted pattern instead. Not checked yet
5 details in Everything
skill-scanner adds OSV.dev OSV.dev An open-source vulnerability database and API maintained by Google that aggregates CVEs and ecosystem-specific advisories in a unified schema, giving cyber tools a single feed to query for package-level vulnerability data. dependency querying, trigger specificity analysis, LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. consensus runs, and token usage reporting
└──▷ USE IT
Enable OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. .dev vulnerable dependency checking and trigger specificity analysis alongside LLM consensus across 3 runs.
yaml
copy
use_osv: true
use_trigger: true
llm_consensus_runs: 3
Adds use_osv config field to query OSV.dev for vulnerable exactly pinned Python dependencies during scans. Not checked yet Adds use_trigger config field to enable trigger specificity analysis. Not checked yet Adds llm_consensus_runs integer config field (default: 1) to control the number of LLM runs used for majority consensus adjudication. Not checked yet Scan results now include aggregate llm_usage with input_tokens, output_tokens, and total_tokens fields when the configured provider reports usage for LLM, meta, or policy-enabled adjudication calls. Not checked yet
4 details in Everything
skill-scanner Python SDK adds llm_usage token reporting and build_analyzers factory for programmatic analyzer assembly.
└──▷ USE IT
Assemble all optional analyzers including behavioral, LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. , and OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. checks with 3 consensus runs — matching what the CLI and API produce.
python
copy
from skill_scanner.core.analyzer_factory import build_analyzers
analyzers = build_analyzers(
policy,
use_behavioral=True,
use_llm=True,
use_osv=True,
llm_consensus_runs=3,
)
scanner = SkillScanner(analyzers=analyzers, policy=policy)
Enable the demote-only adjudication pass before scanning to apply adjudicator logic programmatically.
python
copy
policy.adjudicator.enabled = True
scanner = SkillScanner(analyzers=analyzers, policy=policy)
Adds llm_usage field (dict[str, int]) to scan results, reporting aggregate input, output, and total tokens across LLM, meta, and adjudication calls. Not checked yet Adds build_analyzers factory function in skill_scanner.core.analyzer_factory to assemble optional analyzers (use_behavioral, use_llm, use_osv, llm_consensus_runs) programmatically, matching CLI and API parity. Not checked yet Enables demote-only adjudication pass programmatically via policy.adjudicator.enabled = True before constructing SkillScanner. Not checked yet
3 details in Everything
skill-scanner adds adjudicator and llm_analysis policy keys for false-positive demotion and LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. budget control
└──▷ USE IT
Enable adjudication with a confidence threshold to automatically demote likely false positives while keeping an audit trail.
yaml
copy
adjudicator:
enabled: true
min_fp_confidence: 4
Restrict LLM transitive-trust findings to only your internal registries and package hosts so external references still trigger full-severity alerts.
yaml
copy
llm_analysis:
trusted_reference_domains:
- git.example.com
- packages.example.com
Adds adjudicator.enabled and adjudicator.min_fp_confidence config keys to enable demote-only false-positive review with a configurable confidence threshold; disabled by default and records decisions for audit. Not checked yet Adds llm_analysis.trusted_reference_domains config key to list organization-controlled domains whose transitive-trust and supply-chain findings are automatically demoted to LOW severity. Not checked yet Adds llm_analysis policy section covering prompt/output budgets and a meta multiplier for controlling LLM analysis behavior. Not checked yet
3 details in Everything
skill-scanner gains repo scanning, OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. advisory queries, LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. adjudication, multi-format output, and rule-pack controls in one release.
└──▷ TRY IT
Scan a public GitHub repo for skills, including OSV dependency advisories and overlap checks, without cloning manually.
copy
$ skill-scanner scan-repo https://github.com/owner/repo --use-osv --check-overlap
Run a full local skill scan and emit both a machine-readable JSON report and a SARIF SARIF Static Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. file for upload to a code-scanning platform in one pass.
copy
$ skill-scanner scan ./my-skill --format json --format sarif \
--output-json results.json --output-sarif results.sarif
Review deterministic HIGH/CRITICAL regex matches with LLM adjudication before committing — reduces false-positive noise without a full LLM enrichment pass.
copy
$ skill-scanner scan ./my-skill --adjudicate
New scan-repo subcommand clones a public GitHub repository into a temporary directory, recursively discovers and scans skills, then removes the clone on completion — accepts owner/repo shorthand or a full HTTPS HTTPS HTTP carried over an encrypted TLS connection. It protects traffic in transit and lets a browser verify the server's certificate. URL. Not checked yet Adds --use-osv flag to query OSV.dev OSV.dev An open-source vulnerability database and API maintained by Google that aggregates CVEs and ecosystem-specific advisories in a unified schema, giving cyber tools a single feed to query for package-level vulnerability data. for advisories affecting exactly pinned Python dependencies; no API key required. Not checked yet Adds --adjudicate flag for a demote-only LLM review of deterministic HIGH/CRITICAL literal-regex false positives before LLM enrichment. Not checked yet Adds --llm-max-tokens N to override the LLM response-token budget (default: 8192). Not checked yet Adds --rule-packs PACK... to enable named packs such as atr and promptguard; run --rule-packs list to inspect available packs. Not checked yet
+8 more
show less
Supports repeating --format to produce multiple report formats in a single scan run. Not checked yet Adds per-format output path flags --output-json, --output-sarif, --output-markdown, --output-html, and --output-table. Not checked yet Adds --render-markdown / --no-render-markdown to override terminal Markdown rendering. Not checked yet Adds --taxonomy PATH to override the Cisco AI taxonomy JSON/YAML file. Not checked yet Adds --threat-mapping PATH to override the threat mapping JSON file. Not checked yet Adds openai-compatible as an LLM shortcut value for the LLM provider flag. Not checked yet New skill-scanner-pre-commit hook command with --install to register the built-in Git hook and --scan-all to scan all configured skills; in CI, integrates with pre-commit run skill-scanner --from-ref "$BASE_SHA" --to-ref "$HEAD_SHA" to scope scans to changed paths between two revisions. Not checked yet scan-repo supports --no-recursive to disable recursive skill discovery (recursive is on by default). Not checked yet
13 details in Everything
skill-scanner gains repo scanning, multi-format output, behavioral/LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. analyzers, and rule pack selection in one release.
└──▷ TRY IT
Scan a remote GitHub repository for all skills without cloning it locally — useful in CI pipelines that operate against upstream repos.
copy
$ skill-scanner scan-repo owner/repo
Run the full analyzer stack — behavioral, OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. vulnerability DB, LLM semantic, and metadata — with ATR and PromptGuard rule packs to maximize detection coverage on a local skill.
copy
$ skill-scanner scan /path/to/skill --use-behavioral --use-osv --use-llm --enable-meta --rule-packs atr promptguard
Produce both JSON and SARIF SARIF Static Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. reports from a single scan pass — ideal for feeding results simultaneously into a SIEM SIEM Security information and event management, collecting logs from many systems and correlating events to find suspicious activity. Analysts use it to investigate alerts and incidents. and a code-review tool like GitHub Advanced Security.
copy
$ skill-scanner scan /path/to/skill --format json --format sarif --output-json results.json --output-sarif results.sarif
Adds scan-repo owner/repo subcommand to scan every skill in a public GitHub repository in a single invocation. Not checked yet Adds --use-behavioral, --use-osv, --use-llm, and --enable-meta flags to scan for dependency, OSV, and LLM-powered semantic analysis. Not checked yet Adds --rule-packs to scan and scan-all for selecting optional rule packs; supports atr and promptguard packs, and accepts list as a value to enumerate all bundled packs. Not checked yet Adds --recursive flag to scan-all to traverse nested skill directories. Not checked yet Adds simultaneous multi-format output via --format json --format sarif combined with --output-json <path> and --output-sarif <path>, eliminating the need to scan twice for different report formats. Not checked yet
5 details in Everything
Skill Scanner adds OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. dependency scanning, an adjudicator engine, ATR and PromptGuard rule packs, and extended detection coverage across 9 ecosystems.
└──▷ TRY IT
Run both the ATR and PromptGuard rule packs together to maximise detection coverage in a single scan.
copy
$ skill-scanner scan ./skill --rule-packs atr promptguard
Enable the adjudicator alongside LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. enrichment and the meta-analyzer to reduce false positives before results are promoted.
copy
$ skill-scanner scan ./skill --use-llm --enable-meta --adjudicate
Check pinned Python dependencies against the OSV database without any API key as part of a CI pipeline.
copy
$ skill-scanner scan ./skill --use-osv
Adds --use-osv flag to activate the OSV Analyzer, which checks known vulnerabilities in exactly pinned Python dependencies with no API key required. Not checked yet Adds --adjudicate flag to enable the Adjudicator, a demote-only LLM review that can reduce deterministic HIGH/CRITICAL literal-regex false positives to INFO without ever promoting a finding. Not checked yet Adds --rule-packs atr to enable the ATR (Cisco AI Threat Research) rule pack, currently covering 712 rules. Not checked yet Adds --rule-packs promptguard to enable the PromptGuard rule pack with 26 rules covering PII PII Personally Identifiable Information — data that can identify a specific individual, such as names, email addresses, or SSNs. Cyber tools flag or redact it to limit exposure and meet privacy regulations. , provider secrets, and LLM/Markdown exfiltration. Not checked yet Aggregates Adjudicator, LLM analyzer, and meta-analyzer token usage under llm_usage in JSON output when the provider reports usage. Not checked yet
+4 more
show less
Extends unpinned-dependency detection to Python, Node/npm, Ruby, Go, Rust, Java/Maven, .NET/NuGet NuGet Microsoft's package manager for .NET, hosting a public registry of reusable libraries so tools can pull in dependencies without bundling or manually maintaining third-party code. , PHP/Composer, and Swift manifests. Not checked yet Expands suspicious-URL detection in YAML, JSON, TOML TOML A configuration file format designed to be minimal and readable, mapping unambiguously to a hash table; cyber tools use it as a structured alternative to YAML or INI for config files. , INI, and environment/config files to include bore.pub, serveo.net, and localtunnel.me alongside existing ngrok domains. Not checked yet Adds Unicode Tag Block (ASCII-smuggling) payload detection, generic jailbreak pattern detection, and improved OpenAI/Anthropic project-key detection to core coverage. Not checked yet Adds JavaScript and TypeScript content to analyzability scoring, plus recursive discovery of symlinked skill directories. Not checked yet
9 details in Everything
skill-scanner adds OpenAI-compatible endpoint routing, per-role model overrides, and OSV.dev OSV.dev An open-source vulnerability database and API maintained by Google that aggregates CVEs and ecosystem-specific advisories in a unified schema, giving cyber tools a single feed to query for package-level vulnerability data. dependency scanning via new env vars and --use-osv.
Adds SKILL_SCANNER_LLM_PROVIDER environment variable to route LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. calls, including openai for OpenAI-compatible endpoints. Not checked yet Adds SKILL_SCANNER_LLM_BASE_URL environment variable to point skill-scanner at any OpenAI-compatible base URL. Not checked yet Adds SKILL_SCANNER_LLM_USER environment variable to inject a raw Chat Completions user field on OpenAI-compatible routes. Not checked yet Adds SKILL_SCANNER_LLM_TEMPERATURE environment variable to override sampling temperature, or set to none to omit it entirely. Not checked yet Adds SKILL_SCANNER_ADJUDICATOR_LLM_MODEL environment variable to override the model used specifically for the --adjudicate path (e.g. vertex_ai/gemini-2.5-pro). Not checked yet
+4 more
show less
Supports Vertex AI ambient credentials via GOOGLE_APPLICATION_CREDENTIALS or Application Default Credentials, including Workload Identity, with no extra configuration. Not checked yet Supports ambient AWS/IAM IAM Identity and Access Management — a framework of policies and services for controlling who can authenticate to systems and what resources they are authorized to use, letting cyber tools enforce least-privilege and audit access. credentials for Bedrock with no extra configuration. Not checked yet Adds --use-osv flag for OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. .dev dependency scanning — requires network access but no API key or additional package. Not checked yet GitHub release artifacts now ship with build-provenance provenance Evidence recording where a software artifact came from and how it was built, including its source and build process. It helps a consumer check that the artifact matches a trusted build. attestations and an SBOM SBOM A software bill of materials, an inventory of the components and versions inside an application. It helps teams find affected software when a dependency has a vulnerability or licensing issue. , enabling supply-chain verification and software-inventory retention for controlled environments. Not checked yet
9 details in Everything
Skill Scanner 2.0.13 adds OSV.dev OSV.dev An open-source vulnerability database and API maintained by Google that aggregates CVEs and ecosystem-specific advisories in a unified schema, giving cyber tools a single feed to query for package-level vulnerability data. dependency scanning, unpinned-dependency detection, suspicious-URL classification, trusted-domain LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. policy, and LLM token usage in JSON output.
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.13 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.13
└──▷ TRY IT
Scan a skill's dependencies against the OSV OSV Open Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. advisory database to catch known vulnerabilities in pinned or unpinned packages.
copy
$ skill-scanner scan /path/to/skill --use-osv
Run a full CI scan with LLM analysis and emit SARIF SARIF Static Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. output, while failing the build on any high-severity finding and tracking token spend in the JSON result.
copy
$ skill-scanner scan-all ./skills --recursive --use-llm --use-osv --fail-on-severity high --format sarif --output results.sarif
Adds --use-osv flag to enable the new OSV.dev dependency vulnerability analyzer, checking dependencies against the open OSV advisory database. Not checked yet Adds trusted_reference_domains to the LLM analysis policy to allowlist domains that should not be flagged as suspicious during semantic analysis. Not checked yet Exposes LLM token usage in ScanResult JSON output, including adjudicator token consumption, enabling cost tracking and auditing. Not checked yet Flags unpinned dependencies across manifest formats via the static analyzer, surfacing supply-chain risk in skill packages. Not checked yet Classifies suspicious URLs found in configuration files via the static analyzer. Not checked yet
+4 more
show less
Demotes literal-regex false positives on deterministic HIGH+ findings via the adjudicator, reducing noise without sacrificing real detections. Not checked yet Supports incremental CI skill scans via the pre-commit pre-commit An open-source framework for managing and running git hooks, maintained by the pre-commit project, that lets cyber tools enforce checks automatically on every commit before code reaches a repository. integration, scanning only changed skills instead of the full set. Not checked yet Expands the ATR rule pack from 314 to 712 supported rules (pack version 3.5.6). Not checked yet Attests build provenance provenance Evidence recording where a software artifact came from and how it was built, including its source and build process. It helps a consumer check that the artifact matches a trusted build. and publishes an SBOM SBOM A software bill of materials, an inventory of the components and versions inside an application. It helps teams find affected software when a dependency has a vulnerability or licensing issue. for release artifacts, supporting software supply-chain software supply-chain The dependencies, tools and processes used to build and deliver software. Compromising one part can affect the products or organizations that rely on it. verification. Not checked yet
9 details in Everything
skill-scanner 2.0.12 adds ASCII/Unicode smuggling detection, YARA-based jailbreak rules, and repo-name propagation.
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.12 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.12
Detects ASCII smuggling attacks via Unicode Tag Block characters in scanned skills. Not checked yet Adds a YARA rule for jailbreak detection. Not checked yet Supports passing repository name through the scanner for richer scan context. Not checked yet Supports the OpenAI user request field when using LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. -backed analysis. Not checked yet Allows temperature override and omits temperature for newer Claude and o1 models. Not checked yet
5 details in Everything
skill-scanner 2.0.12 adds ASCII smuggling detection, jailbreak YARA rule, and repo-name propagation for scans.
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.12 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.12
└──▷ TRY IT
Scan a skill for ASCII smuggling and jailbreak patterns — both new detections fire automatically via the static engine, no extra flags needed.
copy
$ skill-scanner scan /path/to/skill --format sarif --output results.sarif
Scan a GitHub repo and surface jailbreak/ASCII-smuggling findings in CI, with the repo name now correctly propagated in scan metadata.
copy
$ skill-scanner scan-repo owner/repo --use-llm --fail-on-severity high --format sarif --output results.sarif
Adds detection of ASCII smuggling via the Unicode Tag Block, catching a covert data-hiding technique used in prompt injection prompt injection An attack where text the model reads, such as a web page, file or tool result, carries instructions that override the user's. It is the main security risk for agents with tools. attacks. Not checked yet Adds a jailbreak detection YARA rule to the static analysis engine. Not checked yet Extends scanner to pass repository name through scan context, improving traceability in multi-repo pipelines. Not checked yet Supports the OpenAI user request field in LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. analyzer calls. Not checked yet Enables temperature override and automatic temperature omission for newer Claude and o1 models in the LLM analyzer. Not checked yet
5 details in Everything
skill-scanner 2.0.10 expands ATR rules 9× to 314, adds --render-markdown flag, and supports OpenAI-compatible endpoints.
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.10 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.10
└──▷ TRY IT
Force markdown rendering on in a terminal that doesn't auto-detect it, useful in CI pipelines where output is piped to a log viewer.
copy
$ skill-scanner scan <target> --render-markdown
Expands the ATR detection pack from 34 to 314 rules, restructured by category. Not checked yet Adds --render-markdown CLI flag to override markdown rendering behavior. Not checked yet Supports OpenAI-compatible LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. endpoints for flexible backend configuration. Not checked yet
3 details in Everything
skill-scanner 2.0.9 adds a PromptGuard rule pack covering PII PII Personally Identifiable Information — data that can identify a specific individual, such as names, email addresses, or SSNs. Cyber tools flag or redact it to limit exposure and meet privacy regulations. , secrets, and LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. exfiltration detection.
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.9 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.9
skill-scanner 2.0.8 adds MCP MCP Model Context Protocol, a standard for connecting AI applications to external tools and data sources. A compatible server exposes capabilities an assistant can use through a common interface. /multi-agent attack detection, a --rule-packs flag, and Anthropic/OpenAI key prefix detection.
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.8 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.8
└──▷ TRY IT
Load the new ATR community rule pack alongside your existing packs to catch MCP tool-poisoning and multi-agent attack patterns in one pass.
copy
$ skill-scanner scan --rule-packs atr,default <target>
Scan an agent codebase for accidentally committed Anthropic or OpenAI project keys as part of a CI pre-merge check.
copy
$ skill-scanner scan --rule-packs atr,default <path>
Adds ATR community rule pack with 34 rules covering MCP tool poisoning, multi-agent attacks, and advanced prompt injection prompt injection An attack where text the model reads, such as a web page, file or tool result, carries instructions that override the user's. It is the main security risk for agents with tools. . Not checked yet New --rule-packs CLI flag enables loading multiple signature rule packs in a single scan invocation. Not checked yet Detects documented Anthropic and OpenAI project API key prefixes to flag exposed credentials. Not checked yet
3 details in Everything
skill-scanner 2.0.7 adds --lenient mode to scan skills that lack a SKILL.md file
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.7 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.7
└──▷ TRY IT
Scan a skill directory that has no SKILL.md — useful for incomplete or in-progress skill packages during development
copy
$ skill-scanner --lenient <path>
skill-scanner 2.0.5 lets --output-<fmt> flags drive the primary output format directly.
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.5 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.5
└──▷ TRY IT
Emit scan results in a specific format as the primary output without extra plumbing.
copy
$ skill-scanner scan <target> --output-json
Supports using --output-<fmt> flags to control the primary output format, not just secondary formats. Not checked yet
1 detail in Everything
skill-scanner 2.0.2 adds GPT-5 support and strengthens multilingual prompt-injection prompt-injection An attack where text the model reads, such as a web page, file or tool result, carries instructions that override the user's. It is the main security risk for agents with tools. detection.
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.2 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.2
Adds GPT-5 model support via drop_params=True for compatibility with the new API. Not checked yet Strengthens multilingual prompt-injection detection guidance for more accurate cross-language findings. Not checked yet
2 details in Everything
skill-scanner 2.0.1 renders Markdown output directly in the terminal via --format markdown.
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.1 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.1
└──▷ TRY IT
Get human-readable, formatted scan results directly in your terminal without piping to a separate Markdown renderer.
copy
$ skill-scanner scan <target> --format markdown
Skill Scanner 2.0 ships a policy engine, JS/TS scanning, five new analysis engines, and an interactive wizard.
└──▷ GET THIS VERSION
copy
$ git clone --branch 2.0.0 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 2.0.0
└──▷ TRY IT
Export a strict baseline policy file, then run a scan against it — useful for enforcing tighter thresholds in CI without touching code.
copy
$ skill-scanner generate-policy --preset strict -o policy.yaml && skill-scanner scan ./my-skill --policy policy.yaml --format sarif --output-sarif results.sarif
Bring your own detection rules for proprietary threat patterns by pointing the scanner at a custom rule pack directory.
copy
$ skill-scanner scan ./my-skill --custom-rules ./packs/my-org-rules/ --format markdown --output-markdown report.md
Launch the interactive wizard for a guided audit — ideal for onboarding new team members or quickly auditing an unfamiliar skill repo.
copy
$ skill-scanner wizard
Adds a YAML-based Scan Policy Engine with three built-in presets (strict, balanced, permissive) and 13 independently configurable sections, replacing all hardcoded thresholds and allowlists. Not checked yet Introduces skill-scanner generate-policy --preset <name> -o policy.yaml to export a baseline policy and skill-scanner configure-policy for a full terminal UI (Textual TUI TUI A text-based user interface that runs inside a terminal, using characters and ANSI escape codes to render menus, panes, and widgets without a graphical desktop, letting cyber tools ship rich interactivity with no GUI dependency. ) to edit every policy setting. Not checked yet Replaces the monolithic signatures.yaml with a modular Rule Pack Architecture (packs/core/) including 14 YARA rule files, 12 Python check modules, a formal pack.yaml manifest, and --custom-rules support for user-supplied packs. Not checked yet Adds first-class JavaScript and TypeScript scanning: eval, child_process, Function constructor, fetch, fs access, secrets, obfuscation patterns, and asset-level prompt-injection prompt-injection An attack where text the model reads, such as a web page, file or tool result, carries instructions that override the user's. It is the main security risk for agents with tools. scanning for .html, .css, .svg, .xml, .xsd. Not checked yet New Strict Structure Validator (SkillValidator) performs pre-load structural checks for hidden files, symlinks, disallowed extensions, binary content, non-UTF-8 encoding, and missing/malformed SKILL.md frontmatter. Not checked yet
+17 more
show less
New Interactive Scan Wizard provides a guided step-by-step scanning mode that auto-detects skills, walks through analyzer/policy/output selection, and generates the equivalent CLI command. Not checked yet New Pipeline Analyzer models taint flow through shell pipelines using a source → transform → sink model to detect multi-step attack chains (data exfiltration data exfiltration Moving data out of a system or organization without authorization. It can happen through network transfers, removable media or other channels. , fetch-execute, find-exec, variable laundering). Not checked yet New Bytecode Analyzer detects .pyc files without matching .py source and bytecode whose AST AST An abstract syntax tree, a structured representation of parsed code. Tools inspect or transform this tree to understand expressions and statements rather than just matching text. doesn't match its source file, flagging hidden code and supply-chain tampering. Not checked yet New Content Extractor recursively unpacks ZIP, TAR, DOCX/XLSX/PPTX, JAR/WAR/APK archives with zip-bomb protection, path-traversal rejection, and VBA macro/OLE detection, producing virtual SkillFile objects for nested analysis. Not checked yet New File Magic Detection (Magika) uses Google Magika (200+ types, ~99% accuracy) with magic-byte fallback to flag extension-vs-content mismatches (e.g., a .py file that is actually an ELF ELF The standard binary format for executables, shared libraries, and object files on Linux and most Unix-like systems, giving cyber tools a predictable structure to parse, instrument, or analyze compiled code. binary) as CRITICAL severity. Not checked yet New Bash Taint Tracker performs lightweight dataflow analysis for bash scripts, tracking taint from command substitution, env vars, sensitive file reads, and network sources through variable propagation to dangerous sinks. Not checked yet Adds a four-tier Command Safety model (SAFE / CAUTION / RISKY / DANGEROUS) that evaluates commands in full context — arguments, pipes, redirects, and chained commands — fully configurable via scan policy. Not checked yet Expands static analysis to a 13-pass architecture covering manifest validation, prompt injection in SKILL.md, Python and Bash code checks, PDF structural analysis, Office macro detection, homoglyph attacks, YARA binary patterns, and asset-level prompt injection. Not checked yet Adds Finding Deduplication: when multiple engines flag the same line for the same threat, findings collapse to a single entry at the highest severity with full engine attribution. Not checked yet Adds Analyzability Scoring (0–100) quantifying what fraction of a skill the scanner can inspect, surfacing a HIGH-severity finding when coverage is low. Not checked yet Introduces two-phase scan execution: Phase 1 runs all deterministic analyzers; Phase 2 feeds their enrichment context to the LLM LLM A large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. for semantic analysis and false-positive filtering. Not checked yet Adds native Azure AD/Entra ID Entra ID Microsoft's cloud-based identity and access management service, formerly Azure Active Directory, used by cyber tools to authenticate users and enforce access policies in Microsoft 365 and Azure environments. authentication for Azure OpenAI via DefaultAzureCredential; install with pip install skill-scanner[azure]. Not checked yet Adds DeepSeek model support with automatic json_object response format fallback for models that don't support JSON Schema mode. Not checked yet Supports multiple simultaneous output formats in a single scan run via repeated --format flags and per-format output flags (--output-json, --output-sarif, --output-markdown, --output-html, --output-table). Not checked yet New standalone HTML Reporter generates self-contained reports with a severity summary, verdict banner, inline pipeline flow diagrams, sortable findings table, and dark theme. Not checked yet Adds --version / -V flag to print the installed version. Not checked yet Adds Homebrew installation support for macOS via brew tap cisco-ai-defense/skill-scanner && brew install skill-scanner. Not checked yet
22 details in Everything
└──▷ BREAKING ON UPGRADE
! The --max-file-size CLI flag is replaced by policy knobs; any automation passing --max-file-size will break on upgrade.! The monolithic signatures.yaml is replaced by the packs/core/ directory with a pack.yaml manifest; any tooling or config referencing signatures.yaml directly will break.
advisory CVE-2026-26057
Feb 17, 2026
·
issue 128
CVE-2026-26057: Skill-scanner Unsecured Network Binding Vulnerability (CVSS CVSS Common Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 6.5, affects <= 1.0.1)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet
skill-scanner 1.0.2 adds custom YARA rules via CLI and retains false-positive metadata in meta-analyzer output.
└──▷ GET THIS VERSION
copy
$ git clone --branch 1.0.2 https://github.com/cisco-ai-defense/skill-scanner.git
$ git checkout 1.0.2
Adds custom rules CLI support, letting practitioners supply their own detection rules at scan time. Not checked yet Retains false positives with metadata in meta-analyzer output, preserving context for triage instead of silently dropping them. Not checked yet Updates AITech Taxonomy with combined threat-classification improvements. Not checked yet
3 details in Everything