SKILLEMALL.ai
GitHub ↗ RU Sign in

CC auditing-mcp-servers-for-tool-poisoning

Audit MCP servers for tool poisoning, tool shadowing, rug pulls, SSRF, and unauthenticated exposure using Invariant Labs' mcp-scan for static/runtime scanning plus manual SSRF/auth checks and description pinning. Use before adding a new MCP server to an agent stack, when reviewing an internal MCP server, detecting rug pulls, or investigating an agent's unexpected tool-driven behavior.

mukul975/Anthropic-Cybersecurity-Skills Agent Skills author: mukul975 Apache-2.0 5 files · 1 script body ≈ 2 499 tokens Open the sourcegithub.com↗ analyzed 5 d ago

Audit MCP servers for tool poisoning, tool shadowing, rug pulls, SSRF, and unauthenticated exposure using Invariant Labs' mcp-scan for static/runtime scanning…

As a process C 53/100 · Has gaps — weak spots: result and completion, when it triggers, failures and branches

AnalyzerVS CodeAI and agentsSecuritytype and topics are labelled automatically from the skill text
Runs in: Claude / Claude Code ChatGPT / Codex Gemini CLI Cursor GitHub Copilot / VS Code Qwen Code Kimi Code CLI Z.ai (GLM) DeepSeek (Deep Code) Mistral Vibe LangChain (deepagents) OpenClaw Goose OpenCode
JSON
Technical rating
C
84/100
safety, quality, tests
Safety 60%
91
Quality 40%
73
Run on models
none yet
Process rating
C
53/100
Has gaps
Result and completion w 14
0
Failures and branches w 10
0
When it triggers w 12
20
the three weakest of ten parameters · all ten

What is at stake

Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.

Obfuscation medium severity

Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.

If you install

The files contain invisible characters, encoded commands or comments hidden from readers but visible to the model. What you read differs from what the agent sees.

For the author

Remove invisible characters (they usually sneak in through copy-paste) and encoded strings: no catalog will pass them. Instructions for the model must be readable by a human too.

How to improve

    For the model run — optional
    • Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
    • A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.

    Guard findings · 5

    ✓ No critical or high findings

    Medium and low: 5
    • medium Obfuscation uni-bidi-override scripts/agent.py:34
      Bidirectional override characters (Trojan Source) (2 occurrences) (detector / deny-list definition)
      SMUGGLE = re.compile(r"[␀-␀␀-␀␀-\U000e0000-\U000e007f]")
      detector
    • low Obfuscation uni-zero-width scripts/agent.py:34
      Zero-width / invisible characters (possible hidden text) (3 occurrences) (detector / deny-list definition; a stray character, not a sequence)
      SMUGGLE = re.compile(r"[␀-␀␀-␀␀-\U000e0000-\U000e007f]")
      detectorstray character
    • low Dangerous commands cmd-pipe-to-shell-known-host scripts/agent.py:70
      Pipe-to-shell installer from a well-known host (still executes remote code) (detector / deny-list definition; string literal in code, not executed)
      "curl -LsSf https://astral.sh/uv/install.sh | sh"}
      detectorcode literal
    • low Dangerous commands cmd-pipe-to-shell-known-host SKILL.md:49
      Pipe-to-shell installer from a well-known host (still executes remote code) (documentation of a security skill)
      curl -LsSf https://astral.sh/uv/install.sh | sh    # or: pipx install uv
      security skill

    A further 1 matches are quotations in this security skill's documentation and are not counted as findings.

    Files scanned: 5. Evidence is masked. Grey chips explain why severity was lowered.

    Against the Agent Skills spec

    • note frontmatter-key unknown frontmatter key "domain"
    • note frontmatter-key unknown frontmatter key "subdomain"
    • note frontmatter-key unknown frontmatter key "nist_ai_rmf"
    • note frontmatter-key unknown frontmatter key "atlas_techniques"

    Process rating: all ten parameters 53/100

    • 0Result and completion. Does not say what the result is
    • 0Failures and branches. Linear process with no failure handling
    • 20When it triggers. No condition that starts the skill
    • 30Running it twice. 1 mutating operations with no state check
    • 60Tools and files. Uses tools (web, python) that frontmatter does not declare
    • 70Inputs and preconditions. Inputs and preconditions are listed
    • 100Steps. 24 steps
    • 100Consistency. Name and required fields are in place
    • 100Execution cost. Instruction body is 2499 tokens
    • 100Progress reporting. Reports progress

    Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.

    Quality signals

    • +5Description has no quoted example phrases that should trigger the skill
    • +4Description does not say when NOT to use the skill (false activations)
    • +3Output format is not stated: the model decides each time
    • -2localhost URLs: will not work for another user
    • -42 reference files, but SKILL.md never points to them: the model will not open them
    • -31 of 1 scripts are never mentioned in SKILL.md
    • +2Single-language instructions
    • +3Description length 387: enough signal without eating the budget
    • +4Structure: 18 headings
    • +3Step-by-step instructions: 24 items
    • +4Has examples (8 code blocks)
    • +1License stated

    Quality base 70; lint remarks subtract, signals add up to 100. Result: 73.