typ-direct-work

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFEPERSISTENCECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PERSISTENCE]: The skill instructs the agent to establish a scheduled task (cron) to handle follow-ups across sessions. While this is a mechanism for persistence, it is used transparently as an operational project management tool.
  • [COMMAND_EXECUTION]: The agent is directed to monitor activity using external tools (e.g., ship ls) and to execute scripts to verify the integrity of project artifacts.
  • [INDIRECT_PROMPT_INJECTION]: The skill involves reading and verifying artifacts produced by external 'crews', which presents a surface for indirect prompt injection if those artifacts contain instructions intended to subvert agent logic. Ingestion points: External crew artifacts. Boundary markers: Not specified in the instructions for artifact processing. Capability inventory: File system access, scheduled task management (cron), and shell command execution. Sanitization: No specific sanitization or validation logic is defined for the content of processed artifacts.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 04:40 PM
Security Audit — agent-trust-hub — typ-direct-work