overbuild-audit
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as project plans, specifications, and code diffs to perform its audit.
- Ingestion points: The skill reads external artifacts including issue descriptions, human anchors in plans, and codebase files (SKILL.md).
- Capability inventory: The agent is authorized to modify the filesystem by deleting code and updating callers when the user explicitly requests to apply the audit's findings (SKILL.md).
- Boundary markers: The skill establishes a logical boundary by requiring the agent to trace all findings back to the original 'ask' in the user's own words, which serves as the authoritative constraint for what should remain.
- Sanitization: No specific sanitization, escaping, or filtering of the external audit targets is described in the instruction files.
Audit Metadata