release
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the project repository to automate documentation and release updates.
- Ingestion points: Reads commit ranges, source code digests, tracked/staged files, READMEs, changelogs, and specifications (SKILL.md).
- Boundary markers: No explicit delimiters or instructions are defined to isolate processed project content from the agent's operational instructions.
- Capability inventory: Performs file system writes for documentation updates and JSON report generation; invokes other internal tools such as
writing,flowmind-author,lev-plan, andcore/eval(SKILL.md). - Sanitization: No sanitization or validation of the ingested project content is mentioned prior to processing or inclusion in generated documentation.
Audit Metadata