bmad-build-auto

Pass

Audited by Gen Agent Trust Hub on Nov 13, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests external project documentation such as tickets and requirements to drive its planning and coding stages.\n
  • Ingestion points: step-01-clarify-and-route.md reads ticket files, epics, and initiative documents.\n
  • Boundary markers: Employs <intent-contract> tags in plan-template.md to isolate read-only instructions.\n
  • Capability inventory: Has file-write permissions, executes shell commands via uv and git, and manages subagents.\n
  • Sanitization: Instructions in edge-case-hunter.md and verification-gap.md restrict subagent behavior to textual analysis and prohibit external tool use.\n- [COMMAND_EXECUTION]: The skill triggers the execution of local project scripts and repository management tools.\n
  • Evidence: SKILL.md directs the user to run uv run for render_skill.py; step-01 and step-03 execute tickets.py and git.\n- [EXTERNAL_DOWNLOADS]: The skill references its remote source code in a GitHub repository managed by the author.\n
  • Evidence: bmod.toml identifies the source as github:bmad-code-org/BMAD-METHOD/skills/method.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 13, 2026, 10:20 PM