claude-devfleet
Pass
Audited by Gen Agent Trust Hub on Jul 4, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to coordinate multiple agents, where the orchestrator ingests mission reports and statuses to manage a project DAG. This architecture presents an indirect prompt injection surface where the behavior of the orchestrator can be influenced by the content of reports generated by other agents.
- Ingestion points: Data enters the orchestrator's context via the
get_reportandget_mission_statustools (SKILL.md). - Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within mission reports.
- Capability inventory: The orchestrator has the ability to dispatch new missions, manage git worktrees, and communicate with a local MCP server (SKILL.md).
- Sanitization: There are no explicit sanitization steps described for the data retrieved from mission reports.
Audit Metadata