gpt-orchestration
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to research codebases and integrate artifacts produced by sub-agents, creating a vulnerability surface where malicious instructions embedded in external files or agent outputs could potentially influence the orchestrator's logic.
- Ingestion points: Processes existing codebases and implementation artifacts from a fleet of agents (SKILL.md).
- Boundary markers: The instructions do not define explicit delimiters or warnings to ignore instructions inside processed data.
- Capability inventory: The skill manages file writes, agent processes, network listeners, and database edges (SKILL.md).
- Sanitization: It instructs the agent to inspect artifacts rather than trusting summaries and use evidence-backed disposition, which serves as a manual validation step but lacks automated sanitization or filtering.
Audit Metadata