yc-office-hours
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the user's project workspace which could contain hidden instructions or malicious content intended to influence the agent's behavior.
- Ingestion points: In Phase 1, the agent is instructed to gather context by inspecting project documentation, existing plans, design documents, and relevant areas of the codebase.
- Boundary markers: The instructions lack explicit boundary markers or warnings to the agent to ignore instructions embedded within the files it reads.
- Capability inventory: The skill has the capability to read from the filesystem and write the final design document back to the project workspace. However, it contains a strict 'Hard Gate' preventing the creation of code, project scaffolding, or any implementation action.
- Sanitization: There is no defined process for sanitizing or validating the content read from the workspace before it is used to inform the agent's strategic diagnostic and recommendations.
Audit Metadata