code-self-describe

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the project's source code and configuration files to generate documentation, which could be exploited to influence agent behavior.
  • Ingestion points: Project source files and dependency manifests such as package.json, go.mod, and Cargo.toml (identified in SKILL.md under Init and Update modes).
  • Boundary markers: The skill uses <!-- auto-generated --> HTML comments to delimit automated sections in CLAUDE.md files, as seen in templates/claude-md-template.md and templates/update-rules.md.
  • Capability inventory: The skill has access to Read, Write, Glob, Grep, Edit, and Bash tools as specified in the SKILL.md frontmatter.
  • Sanitization: The instructions do not specify sanitization or escaping of the content extracted from source files before it is interpolated into documentation.
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to perform project scanning and version control operations.
  • Evidence: Uses git diff --name-only to identify changed files and performs project structure scanning to identify language frameworks and module boundaries.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 08:54 AM