lifecyclemodel-recursive-orchestrator
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/run-lifecyclemodel-recursive-orchestrator.mjsexecutes thetiangong-lcaCLI to perform orchestration tasks such as planning, executing, and publishing model runs. This functionality is restricted to the vendor's own CLI tools and follows standard wrapper patterns. - [INDIRECT_PROMPT_INJECTION]: The skill processes structured request data in JSON format defined by
assets/request.schema.json, which includes fields for notes and metadata. This represents a standard data ingestion surface for the orchestration workflow. - Ingestion points: External JSON request files provided via the
--requestargument, which may contain arbitrary text innotesorgoal.namefields. - Boundary markers: The skill utilizes a strict JSON schema to enforce input structure, though it lacks explicit prompt delimiters to separate data from instructions.
- Capability inventory: The skill can execute shell commands via the
tiangong-lcaCLI launcher. - Sanitization: Inputs are validated against the defined JSON schema and CLI arguments are normalized by the script before execution.
Audit Metadata