docs-manager

Pass

Audited by Gen Agent Trust Hub on Nov 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from repository files (source code, configuration, and documentation) to synchronize documentation state. This behavior presents a vulnerability surface for indirect prompt injection.
  • Ingestion points: Processes repository source code via changed_files, docs/project-config.json, and reference documentation files.
  • Boundary markers: Employs markdown structural delimiters and specific tags like <additional_requests> to isolate inputs.
  • Capability inventory: Performs file system operations, executes local repository scripts via Node.js, and spawns sub-agents.
  • Sanitization: Explicitly requires that paths be passed as separate literal arguments and prohibits raw text interpolation to mitigate command injection risks.
  • [COMMAND_EXECUTION]: The skill facilitates the execution of local repository-based scripts, such as .claude/scripts/doc-impact-map.cjs and .claude/hooks/lib/doc-stamp-guard.cjs, using the node environment to perform logic-heavy tasks.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 4, 2026, 10:12 AM