ship-roadmap

Pass

Audited by Gen Agent Trust Hub on Aug 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references the installation of additional tools via npx skills add gtrabanco/agentic-workflow. This is a vendor-owned resource that extends the skill's capabilities.
  • [COMMAND_EXECUTION]: The skill triggers a local script .agentic-workflow/hooks/fullauto-merge.sh to handle automated merges, allowing user-defined automation for repository state changes.
  • [INDIRECT_PROMPT_INJECTION]: The skill interacts with untrusted data from project roadmaps and issues. It implements an evidence chain to manage this risk: 1. Ingestion points: ROADMAP.md and forge issue trackers. 2. Boundary markers: A locked SHIP_DECISIONS.md file defines the run policy, and GitHub CLI interactions use --body-file to prevent shell injection. 3. Capability inventory: The agent can perform file writes, git operations, and execute the local merge hook as defined in ADVANCE.md. 4. Sanitization: The skill restricts urgent triggers to specific labels and follows a strict founding contract that prevents the agent from seeking new instructions during the loop.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 22, 2026, 01:26 PM
Security Audit — agent-trust-hub — ship-roadmap