prototype
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes content from specification files that may contain instructions for the agent, creating a surface for indirect prompt injection.
- Ingestion points: Untrusted data enters the agent context through [PROTO-RESOLVE: ] markers in Phase 1 specs and user-supplied rationales recorded in decisions.md.
- Boundary markers: Absent. There are no explicit delimiters or instructions provided to the agent to treat the content within these markers as non-executable data.
- Capability inventory: The skill possesses file write (Edit) and shell execution (Bash) capabilities, which could be misused if a malicious instruction is processed.
- Sanitization: Absent. The instructions do not specify any validation or sanitization steps for the ingested strings before they are used in file generation or prompt interpolation.
Audit Metadata