caveman-stats
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads from session logs which contain prior interaction data. This creates a surface for indirect prompt injection if the logs contain adversarial instructions. 1. Ingestion points: JSONL session logs on disk. 2. Boundary markers: Not specified. 3. Capability inventory: Reads session logs, writes a lifetime-savings suffix file, and executes local JavaScript hooks. 4. Sanitization: Not specified.
- [COMMAND_EXECUTION]: The skill depends on external JavaScript hooks (hooks/caveman-stats.js) to intercept and process commands.
- [DATA_EXFILTRATION]: The skill accesses session log files which may contain sensitive conversation history, representing a data exposure risk although no network transmission is explicitly shown in the manifest.
Audit Metadata