mega-code-stop

Pass

Audited by Gen Agent Trust Hub on Mar 20, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: Executes bash commands via uv run to interact with the mega-code CLI. It performs an authentication check and lists active pipelines for user selection.\n- [SAFE]: The workflow incorporates human-in-the-loop confirmation before stopping any processes, mitigating the risk of accidental or malicious termination.\n- [PROMPT_INJECTION]: Potential surface for indirect prompt injection via external tool output.\n
  • Ingestion points: Data enters the context from the output of mega-code pipeline-status in the Workflow section of SKILL.md.\n
  • Boundary markers: No explicit delimiters are used to wrap the tool output.\n
  • Capability inventory: The skill has the Bash tool permitted, which can be used to stop pipelines.\n
  • Sanitization: No programmatic sanitization of run IDs is performed, though the AskUserQuestion tool provides a manual validation step.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 20, 2026, 02:31 AM