teach
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's core functionality involves ingesting and explaining untrusted external data, including source code, git diffs, and PR patches. This creates a surface for indirect prompt injection attacks where an attacker could embed malicious instructions within code comments or documentation intended to bypass the agent's safety constraints or manipulate its output.
- Ingestion points: The skill reads file contents, git history, and patch data into the agent's context through workflows defined in SKILL.md and referenced guides.
- Boundary markers: There are no instructions requiring the use of explicit delimiters or 'ignore embedded instructions' warnings when processing external content.
- Capability inventory: The skill enables the agent to read local files and repository metadata to generate explanations.
- Sanitization: The skill includes robust instructions to redact secrets and private personal data and to exclude sensitive files from analysis, which mitigates data exposure but does not prevent instruction injection from within legitimate project files.
Audit Metadata