devsh-orchestrator
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides a CLI interface (
devsh orchestrate spawn) that executes commands to create and manage remote sub-agents. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and propagate untrusted data (prompts, repository names, and environment variables) to sub-agents.
- Ingestion points: The
--prompt,--repo, and--envflags in thespawncommand. - Boundary markers: None detected in the documentation; it appears to pass prompts directly to the sub-agent interface.
- Capability inventory: The skill can spawn arbitrary agents (
claude/haiku-4.5, etc.), clone repositories, and send inter-agent messages. - Sanitization: Not explicitly mentioned in the documentation; the system relies on the underlying agent platform's safety filters.
Audit Metadata