orchestrate-next-milestone
Pass
Audited by Gen Agent Trust Hub on Mar 28, 2026
Risk Level: SAFEPROMPT_INJECTIONREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill reads milestone plans and style guides from the repository to determine its actions, creating an indirect prompt injection surface. Ingestion points:
docs/active_plans/,refactor_progress.md,AGENTS.md, and style docs. Boundary markers: Not used. Capability inventory: Shell execution (git,pytest,source) and file writing. Sanitization: None observed. - [REMOTE_CODE_EXECUTION]: The skill executes
source source_me.shandpytest, which run code from the local repository. While standard for dev tools, this allows execution of arbitrary code if the repository is untrusted. - [COMMAND_EXECUTION]: The skill executes various shell commands including
git status,git diff, andpytest.
Audit Metadata