orchestrate-omp
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill architecture facilitates the ingestion of data from sub-agents, creating a surface for indirect prompt injection where malicious instructions embedded in sub-agent outputs could influence the coordinator.\n
- Ingestion points: The coordinator is instructed to read results from sub-agents using
agent://<id>andhistory://<id>protocols, as described inSKILL.md.\n - Boundary markers: The instructions recommend structuring shared context using
Goal,Constraints, andContractidentifiers to delimit task information.\n - Capability inventory: The skill utilizes a
tasktool to spawn sub-agents and includes instructions to run validation logic on integrated artifacts.\n - Sanitization: The instructions advise the coordinator to "Inspect each claimed artifact and reconcile conflicts," providing a manual review checkpoint but lacking automated input sanitization.\n- [COMMAND_EXECUTION]: The root coordinator is directed to execute integrated code produced by sub-agents for verification purposes.\n
- Evidence: Section 5 (Integrate) of
SKILL.mddirects the agent to "Run the affected validation and real user path from the root at the integrated head."
Audit Metadata