index-freshness
Warn
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- COMMAND_EXECUTION (MEDIUM): The skill directs the agent to run scripts like ~/.claude-dash/mlx-tools/freshness_checker.py and ~/.claude-dash/watcher/start-watcher.sh. These tools execute with system privileges and accept context-derived parameters.
- PROMPT_INJECTION (LOW): The skill ingests data from index files such as summaries.json and functions.json which could contain adversarial instructions from indexed source code. 1. Ingestion points: summaries.json, functions.json. 2. Boundary markers: Absent. 3. Capability inventory: Subprocess execution of local scripts. 4. Sanitization: Not documented.
Audit Metadata