architecture-survey
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to process and analyze external, untrusted data from the target repository, including source code and commit history, creating an inherent surface for indirect prompt injection.
- Ingestion points: The surveyor is instructed in SKILL.md to read repository files, architectural documentation (CLAUDE.md, ADRs), and the project's commit history to identify candidates for refactoring.
- Boundary markers: The instructions do not define specific delimiters or boundary markers to help the agent distinguish between repository content and its operational instructions.
- Capability inventory: The skill utilizes extensive file system read capabilities and repository metadata analysis; notably, it includes a strict policy against automated code modifications.
- Sanitization: There are no instructions for filtering or sanitizing potentially malicious prompts or instructions embedded within code comments or commit messages.
Audit Metadata