experiment-iterative-coder
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill performs expected software development tasks using well-known, legitimate tools (ruff, pytest) for code quality and verification.
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection, as its primary function involves reading and processing untrusted code and memory files.
- Ingestion points: The skill reads implementation details from workspace files and prior strategies from
/memory/experiment-memory.md(as seen in SKILL.md). - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when the agent reads external code or memory content.
- Capability inventory: The skill has access to shell execution via
execute, as well as file modification tools likewrite_fileandedit_file(defined in allowed-tools). - Sanitization: There is no evidence of sanitization or validation for content ingested from the workspace or memory files before it is processed by the agent.
- Note: This surface is inherent to the skill's intended purpose of code refinement and does not represent malicious intent.
Audit Metadata