retro
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides instructions for a retrospective analysis of coding sessions by reviewing local logs and configuration files. No malicious patterns such as obfuscation, credential harvesting, or unauthorized remote execution were detected.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow to ingest and analyze external data sources such as local session logs and repository configuration files. Ingestion points: session logs and repository metadata files like package.json (SKILL.md). Boundary markers: none present in the instructions. Capability inventory: restricted to reading local files and presenting text-based recommendations to the user. Sanitization: none explicitly defined.
Audit Metadata