agentleFS
Sign inStart free
thedevappsecguythedevappsecguy/sec-skills/skills/skill-scanner-scans/SKILL.md

skill-scanner-scans

Skill0 starsChanged 8 months agoMITView on GitHub

Inspect the current environment, reuse or install the PyPI package `skill-scanner` with `pip`, and scan AI skills, prompts, instructions, agents, and rules for malicious behavior. Use this skill whenever the user mentions `skill-scanner`, `skillscan`, `pip install skill-scanner`, malicious AI skills, suspicious prompt or instruction files, or scanning Codex, Claude, Cursor, Copilot, Gemini, Windsurf, Cline, or OpenCode skill artifacts.

  • Reads credentials
  • Installs packages

The file

Show its rules and text

We split this skill into the rules it gives an agent

32 rules. About 0 are common, 14 are distinctive, and 18 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈14

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 18

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use this skill to set up or reuse the published skill-scanner CLI, discover AI skill and instruction artifacts first, and then scan …

    line 10
  2. Very specific:

    Verify the CLI with skill-scanner --help first and skillscan --help as fallback.

    line 18
  3. Very specific:

    Check analyzer readiness with doctor.

    line 19
  4. Distinctive:

    Run discovery before any live scan.

    line 20
  5. Distinctive:

    If the user already pointed to a specific skill file or folder, scan that target after discovery confirms it.

    line 21
  6. Distinctive:

    If the user did not specify a target, present discovered skills and let the user choose what to scan before running a live scan.

    line 22
  7. Distinctive:

    If no analyzer is configured, still show what is discoverable without keys and stop with the exact blocker.

    line 23
  8. Distinctive:

    Prefer these launcher choices:

    line 30
  9. Very specific:

    Check whether skill-scanner or skillscan is already on PATH.

    line 33
  10. Very specific:

    If the CLI is missing, install with the selected launcher using -m pip install --user skill-scanner.

    line 34
  11. Very specific:

    If the install succeeds but the CLI is not on PATH, derive the user scripts directory from the selected interpreter, run the binary by …

    line 35
  12. Very specific:

    Run skill-scanner doctor after setup.

    line 36
  13. Very specific:

    Run discovery first with skill-scanner discover --format table or skill-scanner scan --list-targets.

    line 37
  14. Very specific:

    If the user provided a specific file or folder, confirm it appears in discovery or is the intended direct target, then run skill-scanner …

    line 38
  15. Distinctive:

    If the user did not provide a target, summarize the discovered skills and ask which one to scan before running the live scan.

    line 39
  16. Distinctive:

    If analyzer configuration does not exist, stop after discovery and report that live scanning is blocked until an analyzer is configured.

    line 40
  17. Distinctive:

    After any live scan, provide clear and actionable remediation for each meaningful finding.

    line 41
  18. Very specific:

    Do not assume uv. This skill intentionally uses pip.

    line 43
  19. Very specific:

    If the user provided a specific file or folder, prefer --path <target> after discovery.

    line 56
  20. Very specific:

    Treat LLM analysis as available only when SKILLSCAN_MODEL is set and either SKILLSCAN_API_KEY or SKILLSCAN_BASE_URL is present.

    line 60
  21. Very specific:

    Treat VirusTotal analysis as available when VT_API_KEY is present.

    line 61
  22. Distinctive:

    Treat live scanning as blocked when neither analyzer is configured.

    line 62
  23. Very specific:

    When live scanning is blocked, continue with discover or scan --list-targets so the user can still see the candidate artifacts.

    line 63
  24. Distinctive:

    Do not skip target selection when multiple discovered skills are in scope and the user did not name one.

    line 64
  25. Distinctive:

    Verify the entrypoint works.

    line 71
  26. Very specific:

    If keys may be valid and the user wants connectivity checked, run doctor --check.

    line 73
  27. Very specific:

    Run discover --format table or scan --list-targets before any live scan.

    line 74
  28. Very specific:

    If the user explicitly named a skill file or folder and at least one analyzer is configured, run scan --path <target> --format summary.

    line 75
  29. Distinctive:

    If discovery returns multiple candidate skills and the user did not specify one, present the list and ask the user which target to scan.

    line 76
  30. Distinctive:

    If discovery returns one obvious target and the user's intent is clearly to scan it, proceed and state that inference.

    line 77
  31. Distinctive:

    If no analyzer is configured, stop after discovery and explain the configuration blocker.

    line 78
  32. Very specific:

    Prefer skill-scanner as the main command name.

    line 80

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Skill Scanner Scans

Overview

Use this skill to set up or reuse the published skill-scanner CLI, discover AI skill and instruction artifacts first, and then scan selected targets for malicious behavior. Prefer direct CLI execution after install, and keep the default output human-readable.

Workflow

  1. Inspect the operating system, Python launchers, Python version, current directory, and whether skill-scanner or skillscan already works.
  2. Reuse an existing working CLI when present.
  3. Select the best Python launcher for the platform and confirm it meets the package requirement of Python 3.11 or newer.
  4. Install with user-level pip only when the CLI is missing or clearly unusable.
  5. Verify the CLI with skill-scanner --help first and skillscan --help as fallback.
  6. Check analyzer readiness with doctor.
  7. Run discovery before any live scan.
  8. If the user already pointed to a specific skill file or folder, scan that target after discovery confirms it.
  9. If the user did not specify a target, present discovered skills and let the user choose what to scan before running a live scan.
  10. If no analyzer is configured, still show what is discoverable without keys and stop with the exact blocker.

Default Behavior

Act instead of only describing:

  1. Detect the platform and available Python launchers before choosing commands.
  2. Prefer these launcher choices:
    • Windows: py -3.11, then py, then python
    • macOS and Linux: python3, then python
  3. Check whether skill-scanner or skillscan is already on PATH.
  4. If the CLI is missing, install with the selected launcher using -m pip install --user skill-scanner.
  5. If the install succeeds but the CLI is not on PATH, derive the user scripts directory from the selected interpreter, run the binary by absolute path, and explain the PATH fix.
  6. Run skill-scanner doctor after setup.
  7. Run discovery first with skill-scanner discover --format table or skill-scanner scan --list-targets.
  8. If the user provided a specific file or folder, confirm it appears in discovery or is the intended direct target, then run skill-scanner scan --path <target> --format summary.
  9. If the user did not provide a target, summarize the discovered skills and ask which one to scan before running the live scan.
  10. If analyzer configuration does not exist, stop after discovery and report that live scanning is blocked until an analyzer is configured.
  11. After any live scan, provide clear and actionable remediation for each meaningful finding.

Do not assume uv. This skill intentionally uses pip.

Environment Checks

Inspect these first:

  • operating system and shell context
  • current directory and whether it is inside a git repository
  • skill-scanner --help
  • skillscan --help
  • Python launcher availability and version
  • whether SKILLSCAN_MODEL, SKILLSCAN_API_KEY, SKILLSCAN_BASE_URL, or VT_API_KEY are set

If the user provided a specific file or folder, prefer --path <target> after discovery. Otherwise, rely on the tool's default broad discovery across repo, user, system, and extension scopes, then let the user choose which discovered skill to scan.

Analyzer Decision

  • Treat LLM analysis as available only when SKILLSCAN_MODEL is set and either SKILLSCAN_API_KEY or SKILLSCAN_BASE_URL is present.
  • Treat VirusTotal analysis as available when VT_API_KEY is present.
  • Treat live scanning as blocked when neither analyzer is configured.
  • When live scanning is blocked, continue with discover or scan --list-targets so the user can still see the candidate artifacts.
  • Do not skip target selection when multiple discovered skills are in scope and the user did not name one.

Default Scan Flow

Use this order:

  1. Reuse or install the CLI.
  2. Verify the entrypoint works.
  3. Run doctor.
  4. If keys may be valid and the user wants connectivity checked, run doctor --check.
  5. Run discover --format table or scan --list-targets before any live scan.
  6. If the user explicitly named a skill file or folder and at least one analyzer is configured, run scan --path <target> --format summary.
  7. If discovery returns multiple candidate skills and the user did not specify one, present the list and ask the user which target to scan.
  8. If discovery returns one obvious target and the user's intent is clearly to scan it, proceed and state that inference.
  9. If no analyzer is configured, stop after discovery and explain the configuration blocker.

Prefer skill-scanner as the main command name. Use skillscan only when the primary entrypoint is unavailable.

Output Pattern

Report progress in this order:

  1. What was found in the environment.
  2. Whether the CLI was reused or installed.
  3. Which Python launcher was selected.
  4. Whether the CLI worked directly or required an absolute scripts-directory path.
  5. Whether analyzers were configured.
  6. Which skills or artifacts were discovered.
  7. Which target was selected and why.
  8. What scan or discovery command was run.
  9. The findings or the exact blocker.
  10. Clear remediation for each real finding, including what to remove, restrict, verify, or rewrite.

Reference Use

Load references/install-and-usage.md when you need:

  • exact Windows, macOS, or Linux install commands
  • user-level scripts-directory and PATH troubleshooting
  • exact doctor, discover, scan, --list-targets, --fail-on, or --output command patterns
  • target-selection and path-scoped scan patterns
  • analyzer environment variables and config file locations

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free