planning-visual-tasks - Decompose visual instructions into structured task plans.
Turns natural-language instructions into structured Plan envelopes of typed sub-goals for host orchestrator execution, with a fallback keyword router.
Tags
Updated: 2026-09-18Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Propose LLM-driven execution plans
- Discover installed tools and models
- Validate plan schemas and DAGs
- Decompose instructions via keyword heuristics
- Emit DAG structure for plans
Inputs
- Natural-language user instruction
- Task state
- Discovered tool inventory
- Plan JSON object
Outputs
- Structured Plan envelope JSON
- Discovered tools JSON object
- Plan validation status JSON
- Execution DAG JSON
Requirements
- Python runtime environment
- Configured LLM provider API key
- Host Orchestrator API environment