afk-workflow

Pass

Audited by Gen Agent Trust Hub on Sep 13, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The workflow is driven by requirements and instructions read from an external ticket tracker ('beads'), creating a surface for indirect injection.
  • Ingestion points: SKILL.md (Step 3) and references/implementer.md instruct the agent to read ticket requirements, acceptance criteria, and notes from the beads tool.
  • Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions when processing external ticket content.
  • Capability inventory: The skill performs file system writes (worktree creation), repository modifications (git commit/merge), and network operations (git fetch/push, npx).
  • Sanitization: There is no mention of sanitizing or validating the content retrieved from the ticket tracker before it influences the agent's implementation or review steps.
  • [EXTERNAL_DOWNLOADS]: The skill explicitly directs the agent to interact with and install external tools from public sources.
  • Evidence: SKILL.md suggests installing mattpocock/skills using the command npx skills@latest add mattpocock/skills.
  • Evidence: The skill relies on the beads project hosted at github.com/gastownhall/beads for state tracking and handoffs.
  • [COMMAND_EXECUTION]: The workflow requires extensive use of shell commands to manage the Git lifecycle and coordination tracking.
  • Evidence: The Merger role (references/merger.md) uses git merge --ff-only and git push with explicit refspecs.
  • Evidence: The workflow utilizes the bd CLI tool for state management (bd update --claim, bd prime).
  • [REMOTE_CODE_EXECUTION]: The suggestion to use npx for installing companion skills involves fetching and executing code from the NPM registry at runtime.
  • Evidence: The 'Companion skills' section in SKILL.md provides a specific command to add external skill packages to the environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 13, 2026, 10:22 PM
Security Audit — agent-trust-hub — afk-workflow