iterate-cr
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes system commands including
git logandgit rev-parseto retrieve repository history, short hashes, and branch names for session tracking. Security constraints are explicitly defined to prevent destructive actions likegit reset,git rebase, or forced pushes. - [INDIRECT_PROMPT_INJECTION]: The skill ingests content from external Change Request documents and previous ledger files, which could influence agent behavior. Ingestion points: Files located at
docs/cr/*.mdanddocs/cr/iterations/*.mdare read into the agent's context during Step 1 and Step 2 of the workflow. Boundary markers: The skill employs a structured markdown template (templates/ITERATE.md) and YAML frontmatter to delimit metadata and entry records. Capability inventory: The agent possesses capabilities to modify project code, write to the filesystem (ledgers), and perform Git commits via the/checkpoint-committool. Sanitization: No explicit content sanitization for the ingested documents is documented, though the process is gated by manual user pacing and identifier validation. - [SAFE]: Repository comparisons and commit links within the changelog point exclusively to the official GitHub organization for the vendor (
github.com/desek), which is consistent with the skill's authorship.
Audit Metadata