gate-work

Pass

Audited by Gen Agent Trust Hub on Jun 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill dynamically parses 'gate' documents and executes shell commands specified within checkpoints to perform verification tasks or build operations.
  • [COMMAND_EXECUTION]: Automates repository lifecycle tasks including 'git commit', 'git push', and modifying version files in response to completed work phases.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing external instructions from workspace-resident files without sanitization.
  • Ingestion points: Files matching '*-gate.md' and 'QUEUE.md' located at the workspace root.
  • Boundary markers: Absent; the instructions direct the agent to 'adopt framing' and work through checkpoints as the primary operational context.
  • Capability inventory: Shell command execution via subprocesses, git repository modification (commit/push), and local file system operations (write, delete, move).
  • Sanitization: None; commands extracted from checkpoints are executed as specified to fulfill validation requirements.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 26, 2026, 04:30 AM
Security Audit — agent-trust-hub — gate-work