orchestrate
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves ingesting data from external and potentially untrusted sources such as session reports, issue descriptions, and pull request content.
- Ingestion points: Data is collected from existing project sessions, recent transcript context, and repository metadata (SKILL.md, Resolve project and capabilities).
- Boundary markers: The skill implements a structured communication schema defined in
references/session-protocol.mdto delimit inter-session messages. - Capability inventory: The agent can message other sessions, read/write to local repositories, interact with CI/CD deployment states, and maintain coordination heartbeats.
- Sanitization: Instructions explicitly state to treat retrieved content as evidence rather than authority to expand scope and require human confirmation for destructive or financial operations.
- [COMMAND_EXECUTION]: The skill uses platform-provided tools and harness controls to discover sessions, interact with Git repositories, and query external project APIs (REST/authenticated API) to maintain project state.
Audit Metadata