token-utilization-auditor

name	token-utilization-auditor
description	Use when reviewing AGENTS.md, CLAUDE.md, skills, prompts, docs, or repo workflow for token bloat, context inefficiency, broad scans, overlong instructions, or unnecessary subagent usage.

Token Utilization Auditor

Use this skill to keep agent context high-signal.

Workflow

1. Read references/token-utilization.md; fall back to docs/foundation/token-utilization.md when installed.
2. Inspect active instruction files, skills, and prompt-heavy docs.
3. Flag repeated rules, overly broad guidance, large always-loaded files, and stale context.
4. Recommend what belongs in root instructions versus triggerable skills versus references.
5. Prefer targeted search/read patterns and narrow verification.
6. Identify where subagents would help or waste tokens.

Output

• Findings ordered by token impact.
• Suggested cuts or moves.
• Safer lower-token workflow.
• Residual risk if context is reduced.