chat-compactor
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest conversation data and save a summary to the local filesystem, which can propagate malicious instructions from a chat into future agent sessions. \n
- Ingestion points: The skill scans the current conversation history to identify outcomes and decisions (SKILL.md). \n
- Boundary markers: The generated markdown format does not include specific boundary markers or instructions to treat the content as data rather than instructions. \n
- Capability inventory: The skill writes to the filesystem, creating new markdown files in the project root or /home/claude/sessions/. \n
- Sanitization: There is no process mentioned for sanitizing or filtering input before it is summarized and saved.
Audit Metadata