tool-call-strategy

What problem does it solve?

This Skill helps agents avoid wasting time, tokens, and context by issuing too many redundant or overly broad tool calls, which can reduce accuracy and make downstream decisions harder.

Core Features & Use Cases

• Three-cost tool discipline: minimizes token overhead, latency, and context pollution while preserving verification quality.
• Script-vs-call decision gate: moves deterministic bulk work into scripts and keeps reasoning-dependent steps in structured tool calls.
• Tool selection decision trees: chooses between file-pattern search, content search, targeted read, and full read to fetch only what is needed.
• Batching and parallelization: combines independent calls into a single round-trip and sequences dependent calls.
• Redundancy avoidance: treats conversation context as a cache to prevent re-reads and re-searches.
• Subagent delegation for context protection: isolates broad exploration to keep the main session clean.

Quick Start

Use the tool-call-strategy skill when your agent is making many tool calls or producing verbose tool output that fills the context window before asking it to complete the task with fewer, better-scoped calls.