vp-long-running-processes
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to inspect output and logs from managed processes to verify health and record status. This creates a surface where instructions embedded in external logs could potentially influence agent behavior.
- Ingestion points: Process logs and command output (SKILL.md).
- Boundary markers: None defined to distinguish between data and instructions in logs.
- Capability inventory: Commands to start, stop, restart, and inspect background processes.
- Sanitization: No specific sanitization mentioned for log content analysis.
- [NO_CODE]: The skill contains no executable scripts, binaries, or source code, consisting solely of markdown instructions and metadata.
Audit Metadata