dev-workflow
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill monitors external issue comments through the
bdt issue watchcommand, creating a surface where untrusted data from third parties enters the agent's context. - Ingestion points: Monitoring issue comments via
bdt issue watch N --follow --skip-own --jsonin SKILL.md. - Boundary markers: No explicit boundary markers or instructions to ignore embedded commands within the comments are provided.
- Capability inventory: The skill has capabilities to perform Git operations, write to the filesystem, update issues, and trigger other agents or sub-agents.
- Sanitization: There is no evidence of sanitization or filtering of the comment content before processing.
- [DYNAMIC_EXECUTION]: The skill utilizes the
codex execcommand to dynamically run code review tasks. - Evidence: Use of
codex exec -C <dir> -s read-only "Use the bms-code-review skill (--full) on the diff ..."in SKILL.md (Step 8) to invoke external logic at runtime.
Audit Metadata