LaunchKit · 2026
Back to Skills

adaptive-reasoning

Automatically assess task complexity and adjust reasoning level.

6
4.2k downloads
by @enzoricciulli

Setup & Installation

openclaw skills install @enzoricciulli/adaptive-reasoning

Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:

npx clawhub install adaptive-reasoning

What This Skill Does

Scores every incoming request on complexity dimensions like multi-step logic, ambiguity, and code architecture, then automatically enables extended thinking when the score warrants it. Appends visual indicators (🧠 or 🧠🔥) to signal when reasoning mode is active. Runs silently as a preprocessing step with no user intervention required.

Avoids the manual overhead of toggling reasoning mode by making the decision automatically based on a consistent scoring rubric.

When to use it

  • Debugging a race condition in a distributed system
  • Designing a caching strategy with conflicting constraints
  • Answering a nuanced trade-off question about system architecture
  • Quickly responding to simple lookups without wasting tokens
  • Auto-downgrading after a complex task when follow-up is trivial

Example Workflow

Here's how your AI assistant might use this skill in practice.

INPUT

User asks: 'Debug why this distributed system has race conditions under load'

AGENT
  1. 1Score the request on complexity dimensions: multi-step logic (+3), code architecture (+2), high stakes (+1) = score 9
  2. 2Threshold ≥8 met, activate extended thinking via /reasoning on
  3. 3Analyze concurrency patterns, shared state, and timing assumptions in the described system
  4. 4Produce a detailed diagnosis with root causes and remediation steps
  5. 5Append 🧠🔥 at the end of the response
OUTPUT

Detailed race condition analysis with identified root causes and fixes, ending with 🧠🔥