define-piece
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the agent to "run it [code, a tool, or a package] in this repository" to verify claims about behavior. This involves executing existing local code to provide evidence for the writing piece.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from multiple project files which could potentially contain malicious instructions. Ingestion points: Repository files including docs/publications/, docs/decisions/, GLOSSARY.md, PRODUCT.md, and source code. Boundary markers: None mentioned. Capability inventory: File writing to docs/briefs/ and GLOSSARY.md; command execution for code verification. Sanitization: None.
- [SAFE]: No malicious patterns such as remote data exfiltration, obfuscation, or unauthorized privilege escalation were identified. The skill's actions are limited to the user's local repository.
Audit Metadata