token-efficient-delegation
Auto-injects token-efficient delegation patterns. Claude tokens are 50x more expensive than MiniMax/GLM/Kimi.
npx skills add majiayu000/claude-skill-registry --skill token-efficient-delegation --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
# Token-Efficient Delegation ## ⛔ CRITICAL: No Claude Subagents **HARD STOP:** Never use `model="haiku"`, `model="sonnet"`, or `model="opus"` in Task tool calls. - Always use: `model` parameter omitted (routes to MiniMax/GLM) - Or explicitly: general-purpose agent (uses cheaper default) - Enforced in: `.claude/settings.local.json` deny rules --- **Core Rule:** Claude = reasoning/decisions, MiniMax/GLM = research/exploration ## The Three Goals (Balance All Three) | Goal | Priority | Strategy | |------|----------|----------| | **Quality** | Keep Opus-level output | Claude synthesizes, subagents gather | | **Cost** | Minimize Claude tokens | Delegate exploration, batch analysis | | **Speed** | Fast execution | Parallel subagents, right model for task | --- ## Quick Decision Tree Before ANY task, ask: 1. **Research/exploration?** → Delegate to MiniMax (fast) or GLM (creative) 2. **Web search needed?** → Delegate to MiniMax 3. **Reading specific known file?** → Claude direct (Read tool) 4. **Analyzing 1-3 images?** → Claude direct 5. **Analyzing 10+ images?** → Parallel subagents (GLM for quality, MiniMax for speed) 6. **Complex reasoning/code?** → Claude direct 7. **Broad codebase sear
- ⛔ CRITICAL: No Claude Subagents
- The Three Goals (Balance All Three)
- Quick Decision Tree
- Model Characteristics (Official Guidance)
- When to Use Each Model
- MiniMax M2.1 - Fast Instruction-Follower
- GLM-4.7 - Creative Problem-Solver
- GLM-4V - Vision Tasks
- Claude (Haiku/Sonnet/Opus) - Reasoning & Decisions
- Quality Preservation Patterns
- Pattern 1: Orchestrator-Worker Split
- Pattern 2: Parallel Research + Single Synthesis
- Pattern 3: Two-Stage Review
- Pattern 4: Specific Prompts
What does the token-efficient-delegation skill do?
Auto-injects token-efficient delegation patterns. Claude tokens are 50x more expensive than MiniMax/GLM/Kimi.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill token-efficient-delegation --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.