ask-pit
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines workflows, such as /triage and /research, that ingest untrusted data from bug reports and external primary sources. This data is subsequently used in downstream tasks like implementation and script generation, creating an attack surface for indirect prompt injection.\n
- Ingestion points: External primary sources via /research, bug reports, and feature requests via /triage.\n
- Boundary markers: The skill emphasizes phase boundaries and context hygiene to manage transitions, though specific technical delimiters for untrusted content are not defined within this router.\n
- Capability inventory: References skills capable of shell script generation (/wizard), file writing (/grill-with-docs, /handoff), and test execution (/tdd).\n
- Sanitization: No explicit sanitization or filtering logic is present in this router skill.\n- [SAFE]: The skill does not contain hardcoded credentials, malicious scripts, or obfuscated content. All referenced external domains, such as GitHub and OpenAI, are well-known services. The skill follows secure practices by directing users to store secrets in .env files rather than hardcoding them.
Audit Metadata