bmad-os-audit-file-refs

Pass

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a shell command sequence involving find, grep, and sort to generate a list of target files for the auditing process.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through the ingestion of external file content.
  • Ingestion points: The skill reads Markdown and YAML files from the src/bmm, src/core, and src/utility directories.
  • Boundary markers: No specific delimiters or boundary markers are used to separate the ingested file content from the subagent's core instructions.
  • Capability inventory: The skill utilizes local shell execution, file system read access, and the ability to spawn and instruct subagents.
  • Sanitization: File content is processed directly by subagents without prior sanitization or instruction-filtering.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 25, 2026, 04:46 AM