Utility Skill for Claude

Utility Skill

Overview

A decision framework for agent orchestration based on Liu et al.,

"Utility-Guided Agent Orchestration for Efficient LLM Tool Use"

(arXiv:2603.19896).

Each candidate action is scored by subtracting weighted costs from

expected gain, producing a single utility value that guides action

selection.

The framework prevents over-calling tools and premature stopping by

making both errors costly.

Utility range is [-2.3, 1.0].

When To Use

• Deciding whether to dispatch another agent or tool call
• Gating expensive tool calls (search, code execution, delegation)
• Selecting the right model tier for a sub-task
• Continuation decisions after receiving partial results
• Verification gating before writing or committing output

When NOT to Use

• Single-step operations with one obvious action
• Trivial tasks where cost of scoring exceeds benefit
• Already-committed actions that cannot be undone

Action Space

A = {respond, retrieve, tool_call, verify, delegate, stop}

| Action | Description |

|-----------|------------------------------------------------------|

| respond | Emit a final answer from current context |

| retrieve | Fetch additional information (search, read, lookup) |

| tool_call | Execute a tool (code runner, API, file write) |

| verify | Check a prior result for correctness or completeness |

| delegate | Spawn a sub-agent or hand off to a specialist |

| stop | Terminate the loop and return current state |

Utility Function

U(a | s_t) = Gain(a | s_t)
           - λ₁ · StepCost(a | s_t)
           - λ₂ · Uncertainty(a | s_t)
           - λ₃ · Redundancy(a | s_t)

| Parameter | Default | Rationale |

|-----------|---------|---------------------------------------------------|

| λ₁ | 1.0 | Cost baseline; all other weights relative to this |

| λ₂ | 0.5 | Weak empirical correlation with outcome (r=0.0131) |

| λ₃ | 0.8 | Redundancy pruning yields ~10% token savings |

Utility range: [-2.3, 1.0].

Positive values indicate the action is worth taking.

Values below the floor (-0.5 default) indicate the action should

be skipped.

Termination Conditions

Stop the loop when any of the following is true:

• (a) Selected action is stop
• (b) Step budget exhausted (default: 10 steps)
• (c) All non-stop actions score below the floor (default: -0.5)

High-gain override: If Gain >= 0.7 for any action, condition

(c) may be overridden.

Document the override and the gain value in your reasoning trace.

Quick Start

Minimal 4-step advisory pattern:

• Construct state: gather task context per

modules/state-builder.md

• Score candidates: evaluate each action in A per

modules/action-selector.md

• Prefer highest utility: select the action with the

maximum U(a | s_t), subject to termination conditions

• Log score and decision: record the winning action,

its utility value, and step count before executing

Detailed Resources

• State Builder: modules/state-builder.md, how to

populate s_t from task context

• Gain: modules/gain.md, estimating expected information

or progress gain

• Step Cost: modules/step-cost.md, token, latency, and

monetary cost tables

• Uncertainty: modules/uncertainty.md, confidence

estimation and calibration

• Redundancy: modules/redundancy.md, detecting duplicate

or low-delta actions

• Action Selector: modules/action-selector.md, scoring

loop and tie-breaking rules

• Integration: modules/integration.md, wiring utility

scoring into existing orchestration loops

Exit Criteria

• [ ] State constructed with task goal and prior steps
• [ ] All six actions scored before selecting one
• [ ] Termination condition checked after each step
• [ ] Score and decision logged for each step taken
• [ ] High-gain overrides documented with gain value