scope-it
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from repository artifacts (ADRs, CONTEXT.md) and tracker items (GitHub issues). Ingestion points:
SKILL.md(Step 1: Orient) andreferences/handoff-inputs.md(Prepare landing obligations). Boundary markers: The skill uses specific comment markers (<!-- scope-it:ticket-map:start -->) for data separation. Capability inventory: Performs file system cleanup, Git operations, and GitHub tracker modifications. Sanitization: Logic in the skill and evals explicitly asserts that external data does not grant invocation authority. - [COMMAND_EXECUTION]: The workflow involves executing commands for repository management, including Git worktree manipulation and file cleanup operations.
- [EXTERNAL_DOWNLOADS]: The skill is distributed via a remote repository and depends on the
to-afk-agentskill for handoff verification.
Audit Metadata