promote
Pass
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses bash commands like grep and sed to locate and search memory files. It interpolates user-provided patterns and keywords directly into shell commands in steps 2 and 6, which presents a potential command injection surface if the platform does not sanitize these inputs.
- [DATA_EXFILTRATION]: Accesses the local project structure, the agent's auto-memory file (MEMORY.md), and configuration files (CLAUDE.md). This data is processed locally to manage rules and is not exfiltrated over the network.
- [PROMPT_INJECTION]: Indirect prompt injection surface identified. 1. Ingestion points: MEMORY.md (Step 2). 2. Boundary markers: Absent. 3. Capability inventory: Shell command execution (grep), file creation/modification (CLAUDE.md, .claude/rules/). 4. Sanitization: Absent; the process relies on agent reasoning and user confirmation steps to validate rules before promotion.
Audit Metadata