swarm
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the agent to construct and execute CLI spawn calls to launch parallel workers. It specifically utilizes the
run_in_background: trueparameter to manage background processes and involves git operations such as worktree management and branch manipulation. - [INDIRECT_PROMPT_INJECTION]: The skill aggregates results from multiple 'worker' roles that ingest untrusted data, including source code, git diffs, and other artifacts, creating a potential path for embedded instructions to influence the parent agent's reporting phase.
- Ingestion points: Terminal results from workers, code artifacts, and git diffs processed in Phase B and Phase C of
SKILL.md. - Boundary markers: The skill lacks explicit boundary markers or instructions to ignore embedded commands; however, it instructs the agent to create a 'compact result table' rather than pasting raw worker output.
- Capability inventory: Spawning background subprocesses via CLI and performing git operations as described in the orchestration steps.
- Sanitization: No specific sanitization or escaping mechanisms are documented for the worker output aggregation process.
Audit Metadata