ai-documentation
Pass
Audited by Gen Agent Trust Hub on Aug 21, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data from a codebase to generate documentation, creating a surface for indirect instructions to influence agent behavior.
- Ingestion points: The agent reads the repository's
README.md,package.json, manifest files, and directory layout (SKILL.md, Investigate section, step 1). It also follows entry points and traces flows in source code (Step 4). - Boundary markers: The instructions do not provide explicit delimiters or warnings to the agent to ignore instructions embedded within the codebase files it is documenting.
- Capability inventory: The skill has access to
Bash,Write,Edit, and the ability to spawn multiplecode-explorersub-agents (SKILL.md, frontmatter and step 3). This combination of reading untrusted content and executing shell commands or sub-agents increases the risk of successful injection. - Sanitization: There is no mention of sanitizing or escaping content read from the files before it is processed by the agent or the sub-agents.
Audit Metadata