bmad-workflow-builder
Warn
Audited by Gen Agent Trust Hub on May 14, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The scripts generate-convert-report.py and generate-html-report.py use subprocess.run with shell=True to open HTML reports on Windows. If the report file path, which includes the skill's name, contains shell metacharacters, it could lead to arbitrary command execution on the host machine.- [REMOTE_CODE_EXECUTION]: The skill's build and conversion processes include an automated step to run unit tests on the resulting scripts. Because the skill can ingest content from arbitrary external URLs for conversion, this creates a vector where a malicious input skill could result in the execution of harmful test code during the build or conversion phase.- [EXTERNAL_DOWNLOADS]: The convert-process.md logic explicitly instructs the agent to fetch content from user-provided URLs. This represents an entry point for untrusted data and potential indirect prompt injection attacks.
Audit Metadata