heartbeat-monitor
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local Node.js scripts located in the plugin's root directory to perform fleet management tasks, such as taking a census of active sessions and validating heartbeat log files.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an ingestion surface for potentially unvetted content by reading data from managed session artifacts.
- Ingestion points: Data is read from heartbeat log files, Git branch heads, pull request metadata, and session worktrees.
- Boundary markers: The instructions explicitly require the agent to quote artifact content verbatim in its rulings, providing clear boundaries for human oversight.
- Capability inventory: The skill uses filesystem access for heartbeat registry management and command execution for operational monitoring.
- Sanitization: A validation script is employed to enforce strict schema requirements on heartbeat logs, rejecting malformed or inconsistent data that could mislead the monitor.
Audit Metadata