bmad-os-audit-file-refs
Pass
Audited by Gen Agent Trust Hub on Apr 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
findcommand to generate a list of markdown and YAML files within specified project directories (src/bmm,src/core,src/utility). This execution is limited to local file discovery and does not involve user-controlled inputs. - [PROMPT_INJECTION]: The skill utilizes subagents to process the content of files from the project source. This creates an indirect prompt injection surface (Category 8) where malicious content within the audited files could attempt to influence the subagent's logic. However, since the subagents are provided with highly specific formatting and reporting instructions, and the operation is central to the skill's auditing purpose, the risk is minimal. There are no explicit boundary markers or sanitization steps used during the interpolation of file content into the subagent prompts.
Audit Metadata