Agent skill
token-economy
Apply token optimization when writing docs, changelogs, MCP tasks. Quality
Install this agent skill to your Project
npx add-skill https://github.com/phamhung075/4genthub-hooks/tree/main/skills/token-economy
SKILL.md
Token Economy
Optimize text for AI context: docs, changelogs, MCP tasks, subtasks.
Principle: Quality first. Never sacrifice essential info.
15 Techniques
| # | Technique | Use | Example |
|---|---|---|---|
| 1 | Tables over prose | Comparisons | Function params → table |
| 2 | Bullets with pipes | Multi-part | Part1 | Part2 | Part3 |
| 3 | Numbered steps | Workflows | 1. Step → 2. Step |
| 4 | One example | Code samples | Keep best, remove rest |
| 5 | Pattern statements | Repetition | **Pattern**: All X do Y |
| 6 | Add "Why" | Decisions | **Why**: [reason] |
| 7 | Concrete errors | Debugging | **Error**: \exact message`` |
| 8 | No fluff | Decorations | Remove ASCII art, emojis |
| 9 | Scannable | Structure | Headers, bullets, tables |
| 10 | Consolidate | Duplicates | Merge similar sections |
| 11 | Compact code | Examples | 3-5 key lines only |
| 12 | Quick lists | Reference | Command/param tables |
| 13 | Inverted pyramid | Info order | Critical → Details → Context |
| 14 | Conditional verbosity | Complexity | High for complex, low for simple |
| 15 | No teaching | Reference | State facts, skip basics |
Quick Workflow
- Identify: Prose → tables? Multiple examples → one? ASCII → numbered? Teaching → facts?
- Apply: Convert, consolidate, compact
- Check: Essential info preserved? Scannable? No fluff?
MCP Templates
Task:
Requirements: [WHAT] | Files: [PATH:LINES] (action) | Acceptance: [CRITERIA] | Why: [CONTEXT]
progress_notes: [ACTION] [WHAT]. [NEXT]
completion_summary: [DONE]. [TECH]. Files: [PATHS]. [IMPACT]
Examples
Before (prose):
This function validates user input. Takes string input_text, returns bool.
If valid returns true, otherwise false. Checks empty, special chars, length ≤100.
After (compact):
**validate_input(input_text: str) → bool**: True if valid (non-empty, no special chars, ≤100)
Before (ASCII art - 180 tokens):
┌─────────┐
│ Auth? │
└────┬────┘
▼
┌─────────┐
│ Valid? │
└──┬───┬──┘
Y N
After (numbered - 36 tokens):
1. Check auth
2. Valid → YES: Process | NO: Reject
Common Fixes
| Issue | Fix |
|---|---|
| Multiple examples | Keep 1 (#4) |
| Prose comparison | Table (#1) |
| Teaching | Facts only (#15) |
| ASCII art | Numbered (#3) |
| No "why" | Add context (#6) |
See TEMPLATES.md for copy-paste templates.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
unused-code-cleanup
Systematically identify and remove unused imports, variables, and dead code from TypeScript/React projects using --noUnusedLocals and --noUnusedParameters compiler flags
spawn-team
Spawn an agent team using the Proxy Pattern. Team lead fetches agent configs from MCP and injects them into teammate prompts (since team agents cannot access MCP tools).
safe-file-removal
Use safe-rm command to safely 'remove' files by renaming them to .obsolete instead of permanent deletion. Reversible, collision-safe, hook-compliant.
changelog-updater
Update CHANGELOG.md and TEST-CHANGELOG.md with new entries following Keep a Changelog format and token optimization principles. Use when adding changes to the changelog, documenting new features, fixes, or optimizations.
handoff
Compact the current conversation into a handoff document for another agent to pick up.
obsidian-vault
Search, create, and manage notes in the Obsidian vault with wikilinks and index notes. Use when user wants to find, create, or organize notes in Obsidian.
Didn't find tool you were looking for?