implement-all
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data like specifications and design documents to drive automated actions, presenting a surface for indirect instructions.
- Ingestion points: Reads specifications, tickets, and design.md from the environment (SKILL.md, Step 1).
- Boundary markers: No specific delimiters are used to separate user instructions from potentially malicious content in the documents.
- Capability inventory: Can execute git, gh, and glab commands, manage branches, and orchestrate sub-agents.
- Sanitization: Lacks explicit validation or sanitization of input documents before using them to direct agent actions.
Audit Metadata