working-on-issues
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements engineering workflow logic using established CLI tools including
gitandgh(GitHub CLI). These operations (e.g.,git worktree list,git fetch,gh issue develop) are consistent with the skill's stated purpose of managing development sessions. - [INDIRECT_PROMPT_INJECTION]: The skill features an attack surface for indirect prompt injection by fetching and processing external data such as issue titles, IDs, and metadata from GitHub and Linear.
- Ingestion points: Data is retrieved from live trackers (GitHub/Linear) and the current repository state (SKILL.md).
- Boundary markers: The instructions specify using native relationships and explicit identifiers, but lack explicit delimiters for variable interpolation.
- Capability inventory: The agent can perform file system operations via
git(checkout, branch creation, worktree management) and network operations via theghtool. - Sanitization: No explicit sanitization or instruction-ignoring boundaries are defined for the fetched issue content. However, the rigid procedural 'gates' defined in the instructions serve as a functional control against arbitrary instruction following.
Audit Metadata