spec-mining
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted data from an external codebase, including source files, tests, and migration histories, which presents a surface for indirect prompt injection if those files contain malicious instructions meant to influence the agent.
- Ingestion points: Explicitly instructs the agent to read the repository's source code, test files, and database migration history (SKILL.md).
- Boundary markers: Absent. The instructions do not define delimiters or provide warnings to the agent to ignore instructions embedded within the code being analyzed.
- Capability inventory: The skill performs file system write operations to save documentation to the
specs/directory. - Sanitization: Absent. The skill does not specify any sanitization, filtering, or validation of the data read from the project files before processing it.
- [SAFE]: The skill's primary purpose is documentation and requirements extraction. It does not perform network operations, execute shell commands outside of reading files, or attempt to escalate privileges.
Audit Metadata