sus-spec
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the user's workspace, which is a standard attack surface for development tools.
- Ingestion points: Reading project rules, code, durable decisions, and open issues as specified in the 'Method' section.
- Boundary markers: The skill does not specify markers to delimit the untrusted workspace data from its internal logic.
- Capability inventory: The skill has capabilities to write files to a local directory and delete files during the 'Close' phase.
- Sanitization: No explicit sanitization or filtering of workspace content is mentioned.
- [SAFE]: The skill implements significant safeguards by explicitly requiring human approval for file operations, including 'Delete', 'Leave', or 'Promote' choices, and strictly forbidding automatic file overwrites.
Audit Metadata