pk-init

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses a shell command (mkdir -p) to initialize the .cockpit/ directory structure. This is a standard setup procedure restricted to creating local folders.
  • [DATA_EXPOSURE]: The skill reads and writes project-specific files like objective.md, tools.md, and AGENTS.md. No sensitive system files, environment variables, or hardcoded credentials were detected.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user input to define project goals and tools. It includes a built-in safety mechanism requiring a 'confirm table before write' step, which ensures human-in-the-loop validation of generated content.
  • [REMOTE_CODE_EXECUTION]: No remote downloads, script piping, or external execution patterns were found. All content is generated locally based on user input and internal templates.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 07:34 AM
Security Audit — agent-trust-hub — pk-init