memory-prune
Pass
Audited by Gen Agent Trust Hub on Aug 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes markdown files that may contain instructions designed to manipulate the agent's behavior during the pruning process.\n
- Ingestion points: Markdown files (*.md) located in ~/.claude/projects//memory/ (SKILL.md).\n
- Boundary markers: None explicitly defined in the instructions to separate data from commands.\n
- Capability inventory: File system write and delete access, as well as the ability to load other skill tools (SKILL.md).\n
- Sanitization: No specific sanitization or filtering of input content is mentioned.\n
- Mitigation: A manual confirmation step is required before the agent applies any changes in Step 3 (SKILL.md).
Audit Metadata