Updated today
by cisco-ai-defense
skill-scanner
Security Scanner for Agent Skills
2.6k GitHub stars (2,605)
+50 in 20 days
| Period | New stars |
|---|---|
| Sep 27 | +1 |
| Sep 28 | +1 |
| Sep 29 | +7 |
| Sep 30 | +5 |
| Oct 1 | 0 |
| Oct 2 | +1 |
| Oct 3 | +2 |
| Oct 4 | 0 |
| Oct 5 | 0 |
| Oct 6 | +2 |
| Oct 7 | +4 |
| Oct 8 | +3 |
| Oct 9 | +2 |
| Oct 10 | +4 |
| Oct 11 | +2 |
| Oct 12 | +3 |
| Oct 13 | +2 |
| Oct 14 | +5 |
| Oct 15 | +2 |
| Oct 16 | +4 |
- 30 skills (30 skills found in its SKILL.md files)
- 332 forks (332 forks)
- Today last update (Last updated today)
Install
Skills CLI
npx skills add cisco-ai-defense/skill-scanner
Asks which skills and agents to install. Just one: npx skills add cisco-ai-defense/skill-scanner --skill no-confirmation-agent — add -g for every project.
Claude Code
git clone https://github.com/cisco-ai-defense/skill-scannercp -r "skill-scanner/evals/skills/autonomy-abuse/no-confirmation" ~/.claude/skills/
Copies the no-confirmation-agent skill into ~/.claude/skills/ for every project (use .claude/skills/ for one project). Repeat for the other skills listed below.
Plugin
/plugin marketplace add cisco-ai-defense/skill-scanner
Run inside Claude Code if the repository ships a plugin marketplace, then pick plugins from /plugin.
Skills can run code on your machine — read the SKILL.md first.
What it does
In the author's own words, from each skill's SKILL.md — showing the first 30 of 33 files.
-
no-confirmation-agent
Direct the agent to act while bypassing user confirmation
evals/skills/autonomy-abuse/no-confirmation/SKILL.md on GitHub (opens in a new tab) -
dynamic-code-compiler
Compile caller-provided source text into an executable code object
evals/skills/backdoor/magic-string-trigger/SKILL.md on GitHub (opens in a new tab) -
external-socket-reporter
Send a diagnostic record over an explicitly external raw socket
evals/skills/behavioral-analysis/multi-file-exfiltration/SKILL.md on GitHub (opens in a new tab) -
expression-compiler
Compile a caller-provided expression into a Python code object
evals/skills/command-injection/eval-execution/SKILL.md on GitHub (opens in a new tab) -
calendar-sync-helper
Sync calendar events with an endpoint selected in configuration
evals/skills/data-exfiltration/config-tunnel-exfil/SKILL.md on GitHub (opens in a new tab) -
sensitive-file-reader
Read a sensitive operating-system credential file for diagnostics
evals/skills/data-exfiltration/environment-secrets/SKILL.md on GitHub (opens in a new tab) -
embedded-payment-key
Demonstrate a credential-shaped payment key embedded in instructions
evals/skills/hardcoded-secrets/stripe-key/SKILL.md on GitHub (opens in a new tab) -
ransomware-request
Issue an explicit imperative request for a prohibited cyber-harm capability
evals/skills/harmful-content/explicit-ransomware-request/SKILL.md on GitHub (opens in a new tab) -
ransomware-chain
Demonstrate a multi-stage file encryption and extortion implementation
evals/skills/malware/ransomware-chain/SKILL.md on GitHub (opens in a new tab) -
xor-code-compiler
Transform an integer with XOR and compile caller-provided source
evals/skills/obfuscation/base64-payload/SKILL.md on GitHub (opens in a new tab) -
unconstrained-file-reader
Read a caller-selected path without confining it to a safe root
evals/skills/path-traversal/file-reader/SKILL.md on GitHub (opens in a new tab) -
unlicensed-helper
Perform a benign transformation while omitting license metadata
evals/skills/policy-violation/missing-license/SKILL.md on GitHub (opens in a new tab) -
concealed-action
Conceal an agent action from the user who requested the task
evals/skills/prompt-injection/jailbreak-override/SKILL.md on GitHub (opens in a new tab) -
unbounded-data-analyzer
Analyze an input repeatedly without a termination condition
evals/skills/resource-exhaustion/infinite-loop/SKILL.md on GitHub (opens in a new tab) -
file-validator
Validate a filename extension without opening any filesystem path
evals/skills/safe-skills-2/file-validator/SKILL.md on GitHub (opens in a new tab) -
simple-math
Perform basic arithmetic using a direct operator implementation
evals/skills/safe-skills/simple-math/SKILL.md on GitHub (opens in a new tab) -
keyword-stuffing
Inflate skill discovery through repeated activation keywords
evals/skills/skill-discovery-abuse/keyword-stuffing/SKILL.md on GitHub (opens in a new tab) -
vague-utility
Utility
evals/skills/social-engineering/vague-description/SKILL.md on GitHub (opens in a new tab) -
interpolated-database-query
Construct a database query by interpolating a caller-controlled identifier
evals/skills/sql-injection/database-query/SKILL.md on GitHub (opens in a new tab) -
unpinned-dependency
Demonstrate a package dependency that is not locked to an exact version
evals/skills/supply-chain-attack/unpinned-dependency/SKILL.md on GitHub (opens in a new tab) -
attacker-forwarding
Forward collected data to an explicitly attacker-controlled destination
evals/skills/tool-chaining-abuse/attacker-forwarding/SKILL.md on GitHub (opens in a new tab) -
external-instruction-follower
Delegate agent behavior to instructions obtained from an external page
evals/skills/transitive-trust-abuse/external-instructions/SKILL.md on GitHub (opens in a new tab) -
remote-package-installer
Install a Python package directly from an untrusted remote archive
evals/skills/unauthorized-tool-use/remote-package-source/SKILL.md on GitHub (opens in a new tab) -
encoded-tag-marker
Contain an encoded Unicode Tag Block marker used for hidden instructions
evals/skills/unicode-steganography/encoded-tag/SKILL.md on GitHub (opens in a new tab) -
weather-assistant
Retrieves current weather conditions and forecasts for any location worldwide
evals/test_skills/malicious/ascii-smuggling/SKILL.md on GitHub (opens in a new tab) -
eicar-test
Test skill containing EICAR test file for malware detection
evals/test_skills/malicious/eicar-test/SKILL.md on GitHub (opens in a new tab) -
data-exfiltrator
Analyzes data files
evals/test_skills/malicious/exfiltrator/SKILL.md on GitHub (opens in a new tab) -
flowise-cve-2025-59528
No description in the SKILL.md.
evals/test_skills/malicious/flowise-cve-2025-59528/SKILL.md on GitHub (opens in a new tab) -
mcp-atlassian-cve-2026-27825
No description in the SKILL.md.
evals/test_skills/malicious/mcp-atlassian-cve-2026-27825/SKILL.md on GitHub (opens in a new tab) -
mcpwn-cve-2026-33032
No description in the SKILL.md.
evals/test_skills/malicious/mcpwn-cve-2026-33032/SKILL.md on GitHub (opens in a new tab)
No skill matches that filter.
Review before you install
A skill is a set of instructions and scripts your agent will follow, and it can run code on your machine. Read the SKILL.md and any scripts first — this listing is not a security review.