Skillv1.0.0

ClawScan security

Markdown Linter · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

BenignMar 11, 2026, 2:07 PM

Verdict: benign
Confidence: medium
Model: gpt-5-mini
Summary: The skill is internally consistent: it lints Markdown files, requires python3 and reasonable Python packages, and only optionally makes network requests when external-link checking is enabled.
Guidance: This linter appears to do what it says: it reads Markdown files you point it at and optionally makes HTTP requests to check external links. Before installing or running it, consider: (1) Don’t run it with broad globs ("**/*.md") on untrusted repositories if you’re worried about scanning many files you didn’t intend to expose. (2) Avoid enabling --check-external-links on untrusted input because the tool will make outbound requests to URLs found in the documents (this could be abused to probe internal services or contact attacker-controlled endpoints). (3) Review the full scripts/main.py (the provided snippet was truncated) yourself if you want higher assurance that there are no hidden network endpoints or suspicious behavior. If you need to be extra cautious, run the linter in a sandboxed environment or inspect/execute it locally rather than allowing autonomous agent invocation.

Purpose & Capability: okName/description match the provided artifacts. The SKILL.md declares python3 and Python packages (markdown, requests) which are appropriate for a Markdown linter that optionally checks external links. The included script implements linting rules that align with the claimed features.
Instruction Scope: noteRuntime instructions tell the agent to run scripts/main.py against one or more Markdown files (glob patterns supported). That necessarily reads the target files (expected), and — if --check-external-links is set — performs HTTP requests to URLs found in the documents. Be aware that running with wide globs (e.g., "**/*.md") will cause the tool to open many files in the workspace and enabling external-link checking will make outbound HTTP requests to whichever URLs appear in those files.
Install Mechanism: okNo install spec is provided (instruction-only install); the skill ships a Python script and expects runtime dependencies to be available. Nothing is downloaded from third-party URLs during install.
Credentials: okThe skill does not request environment variables, credentials, or config paths. The only potentially sensitive operation is optional network access (requests) when external link checking is enabled; no secrets are required.
Persistence & Privilege: okalways is false and the skill does not request persistent/system-wide privileges. There is no indication it modifies other skills or global agent config.