docs
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external, potentially untrusted data from a repository, including source code, configuration files, and existing documentation. This creates an attack surface where malicious instructions embedded in the analyzed code (e.g., in comments or documentation blocks) could attempt to influence the agent's behavior.
- Ingestion points: The skill explicitly directs the agent to read repository files in
SKILL.md("build evidence from the repo"),reference/editor.md("Read the code, tests, scripts, config"), andreference/architecture.md("Find the runtime entrypoints and execution boundaries"). - Boundary markers: The skill provides structural boundaries via the "Handoff" section in
SKILL.mdand specific "Completion criteria," which define the expected output format and verification steps. - Capability inventory: The skill is authorized to read filesystem content and write/modify documentation files within the repository context.
- Sanitization: The skill includes manual sanitization rules, explicitly instructing the agent to exclude "secrets, credentials, raw tokens, or secret values" and to avoid using absolute filesystem paths that might reveal local identity details.
Audit Metadata