prd-to-prod-autopilot

Pass

Audited by Gen Agent Trust Hub on Jul 8, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill relies on external 'Sources of Truth' such as PRDs, issue trackers, and existing project documentation. This creates an indirect prompt injection surface where malicious instructions embedded in a PRD or issue could potentially override the conductor's logic or influence the sub-agents it launches.
  • Ingestion points: PRD source, issue tracker descriptions, and repository instructions.
  • Boundary markers: The skill does not specify the use of delimiters or 'ignore' instructions when interpolating these sources into agent prompts.
  • Capability inventory: The skill can create issues, modify worktrees, push code to GitHub, and launch autonomous sub-agents.
  • Sanitization: No specific sanitization or validation of input data is documented.
  • [COMMAND_EXECUTION]: The skill orchestrates 'verification commands' and implementation workflows. This involves executing shell commands to run tests and verify acceptance criteria, which is a necessary but powerful capability for an autopilot developer tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 8, 2026, 06:47 AM
Security Audit — agent-trust-hub — prd-to-prod-autopilot