runtime-monitor
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines monitoring procedures for backend infrastructure (Redis, Restate, Dkron, Inngest) using internal vendor tools. It includes safety rules for terminal handling to prevent UI crashes.
- [COMMAND_EXECUTION]: Utilizes the
joelclawCLI tool for querying job status, queue depth, and substrate health. These commands are consistent with the developer infrastructure (joelhooks) and are used for read-only monitoring. - [INDIRECT_PROMPT_INJECTION]: The skill ingests workload status data from internal system commands to populate dashboards.
- Ingestion points: CLI output from
joelclaw jobs statusand theruntime_jobs_monitorextension inSKILL.md. - Boundary markers: Not explicitly defined for the command outputs.
- Capability inventory: Limited to TUI updates, status reporting, and OTEL emission; no arbitrary file writing or external network exfiltration capabilities are exposed by the monitoring logic.
- Sanitization: None explicitly documented, though the skill focuses on structured status aggregation.
Audit Metadata