Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
A deterministic rule-based component that classifies clarified user input into exactly one of 15 predefined task categories. It also assigns complexity and risk levels, computes a confidence score, and sets the state transition before task decomposition begins. Runs inside a multi-agent pipeline between the clarification and decomposition stages.
Deterministic tie-breaking rules with a fixed priority order eliminate the ambiguous or inconsistent routing that probabilistic classifiers produce under edge cases.
When to use it
- Routing a 'fix this error' request to a debugging handler instead of a code generation one
- Separating refactoring tasks from feature addition requests before planning
- Escalating high-risk deployment requests for review before execution
- Assigning complexity tiers to incoming requests to select the appropriate execution path
- Distinguishing data analysis jobs from general concept analysis in an agent queue
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Debug the authentication error and update the related docs
- 1Verify input is marked as clarified and has no pending questions
- 2Count keyword matches across candidate categories: DEBUGGING (debug, error), DOCUMENTATION (update, docs)
- 3Apply primary action verb analysis: 'Debug' maps to DEBUGGING
- 4Assign primary category DEBUGGING, secondary tag [DOCUMENTATION]
- 5Compute confidence score, assess risk level, and set state transition to decomposition
ClassificationResult with primary=DEBUGGING, secondary_tags=[DOCUMENTATION], complexity and risk levels assigned, ready for task decomposition