dispatch-issues
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell commands via git to manage branches and worktrees. This includes identifying the current repository state with git status and creating isolated directory structures for subagents within the .worktrees path.
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by ingesting data from external issue trackers and local repository documentation. 1. Ingestion points: External tracker issues and documentation files like AGENTS.md and README.md. 2. Boundary markers: The subagent prompt lacks explicit delimiters for instructions in issue data. 3. Capability inventory: The skill spawns subagents and executes git commands. 4. Sanitization: No explicit sanitization of ingested issue content. The skill mitigates risks by enforcing selection criteria that exclude ambiguous or sensitive tasks.
Audit Metadata