skill CX Codex

skill-optimizer — Claude Code skill: tune skills, run benchmarks

Q: How do I install skill-optimizer in Claude?

Run `npx skills add mcollina/skills --skill skill-optimizer -a claude-code`. The skill file lands at `~/.claude/skills/skill-optimizer/SKILL.md`.

A skill for people who write and edit skills: it shows why a model ignores yours and whether it makes output worse.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓

npx skills add mcollina/skills --skill skill-optimizer -a claude-code

Open on GitHub

What skill-optimizer does

skill-optimizer runs an optimization loop over your skills. You measure baseline and skill-on behavior per model, per scenario, per criterion. Then you classify the failure: the skill never fires, one model lags, or the delta goes negative. Edits target salience — explicit triggers, concrete integrated examples, tighter checklists and decision rules. Re-run the evals, compare deltas, ship with a documented gate, run history and follow-up issues.

Five rule files carry the detail: rules/benchmark-loop.md for the end-to-end loop and scoring, rules/activation-design.md for retrieval and instruction uptake, rules/context-budget.md for cutting token cost, rules/regression-triage.md for diagnosing regressions and rules/release-gates.md for go/no-go criteria. The heuristics are blunt: a few high-signal rules beat many soft recommendations, fragile behavior belongs in top-level checklists, every common scenario needs at least one integrated example, and what must not be omitted needs its own wording.

The skill does not run evals for you. Scenarios and scoring are your job; it supplies the procedure and templates. Without a way to compare output with and without the skill, the loop degrades into guessing at wording. Set up a repeatable run first, then touch the text.

What it helps with

find out why a model ignores your skill
separate a universal failure from one model's weakness
catch a regression caused by the skill itself
shrink skill context without losing behavior
set a release gate for skill updates

Example prompts for the agent

run the skill benchmark and show me the with-skill vs without-skill delta
why does this criterion fail across every model?
cut the skill's context cost and check the behavior still holds

How to install skill-optimizer

1Skills CLI (Vercel)

npx skills add mcollina/skills --skill skill-optimizer -a claude-code

2Manually: put the files in place

~/.claude/skills/skill-optimizer/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/skill-optimizer/

1Skills CLI (Vercel)

npx skills add mcollina/skills --skill skill-optimizer -a codex

2Manually: put the files in place

.agents/skills/skill-optimizer/SKILL.md

For all projects use ~/.agents/skills/

1Skills CLI (Vercel)

npx skills add mcollina/skills --skill skill-optimizer -a cursor

2Manually: put the files in place

.cursor/skills/skill-optimizer/SKILL.md

Cursor also reads .agents/skills/ and .claude/skills/

1GitHub CLI

gh skill install mcollina/skills skill-optimizer

2Manually: put the files in place

.github/skills/skill-optimizer/SKILL.md

1Skills CLI (Vercel)

npx skills add mcollina/skills --skill skill-optimizer -a gemini-cli

2Platform CLI

gemini skills install https://github.com/mcollina/skills --path skills/skill-optimizer

Manually: put the files in place

.devin/skills/skill-optimizer/SKILL.md

The legacy .windsurf/skills/ path still works

Skills CLI (Vercel)

npx skills add mcollina/skills --skill skill-optimizer -a opencode

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about skill-optimizer

How do I install skill-optimizer in Claude?

Run npx skills add mcollina/skills --skill skill-optimizer -a claude-code. The skill file lands at ~/.claude/skills/skill-optimizer/SKILL.md.

What does skill-optimizer do?

It lays out a loop for improving skills: measure baseline and skill-on behavior, classify the failure pattern, edit for salience, re-run evals and ship behind a gate. Five rule files hold the procedures and templates.

Do I need my own evals for it to work?

Yes. The skill provides the procedure and templates; you build the scenarios and scoring. Without a with-skill vs without-skill comparison, edits to the text are guesswork.

Source text

SKILL.md · MIT

---
name: skill-optimizer
description: "Optimizes AI skills for activation, clarity, and cross-model reliability. Use when creating or editing skill packs, diagnosing weak skill uptake, reducing regressions, tuning instruction salience, improving examples, shrinking context cost, or setting benchmark/release gates for skills. Trigger terms: skill optimization, activation gap, benchmark skill, with/without skill delta, regression, context budget, prompt salience."
metadata:
  tags: skills, optimization, benchmarking, activation, regressions, prompt-engineering
---

## When to use

Use this skill when you need to:
- Improve whether a skill is actually applied by models
- Diagnose why some criteria fail across all models
- Prevent a skill from making outputs worse
- Refactor skill text for stronger retrieval under context pressure
- Build repeatable benchmark loops and release gates

## Optimization loop (default workflow)

1. **Measure baseline and skill-on behavior** (per model, per scenario, per criterion)
2. **Find failure pattern**:
   - universal failure (0% with skill)
   - model-specific weakness
   - regression (negative delta)
3. **Edit for salience**:
   - add explicit triggers
   - add concrete integrated examples
   - tighten checklists and decision rules
4. **Re-run evals** and compare deltas
5. **Ship with guardrails** (documented gate + run history + follow-up issues)

## How to use

Read individual rule files for detailed procedures and templates:

- [rules/benchmark-loop.md](rules/benchmark-loop.md) - End-to-end benchmark loop and scoring
- [rules/activation-design.md](rules/activation-design.md) - Improve retrieval and instruction uptake
- [rules/context-budget.md](rules/context-budget.md) - Reduce token cost without losing behavior
- [rules/regression-triage.md](rules/regression-triage.md) - Diagnose and fix skill-on regressions
- [rules/release-gates.md](rules/release-gates.md) - Go/no-go criteria before shipping skill updates

## Practical heuristics

- Prefer **few high-signal rules** over many soft recommendations
- Put fragile, high-value behaviors in **top-level checklists**
- Include at least one **integrated example** per common scenario
- Add explicit wording for what must **not** be omitted
- Track gains/losses with **with-skill vs without-skill** comparisons

prompt-optimizer getsentry A skill that rewrites system and agent prompts against a set of eval cases instead of gut feeling. Useful when a prompt exists but behaves inconsistently. CX Codex +1 2,300installs senior-prompt-engineer Alireza Rezvani Skill for systematic prompt optimization via eval sets, RAG quality measurement, and agent validation. Works with any model, uses only stdlib Python. 970installs bmad-eval bmad-code-org Runs automated benchmarks and evaluations for agent skills across four modes, checking trigger accuracy and grading quality in a clean workspace. CX Codex +1 4,600installs principle-explain-the-number backnotprop A skill for the moment before you trust a measured number: name what caps it, then rule out that the test measured something other than your work. CX Codex +1 4,100installs write-judge-prompt ai-evals-course Walks you through one binary Pass/Fail LLM-as-Judge evaluator for a subjective failure mode: tone, faithfulness, relevance. Useful when code checks fall short. CX Codex +1 1,100installs prompt-governance Alireza Rezvani Manages production prompts as versioned code, builds automated eval pipelines, and prevents quality regressions across AI deployments. 40installs

From the same set11

typescript-magician Resolves TypeScript compiler errors, eliminates any types, and creates strict generic patterns. Useful when locking down codebase type safety. CX Codex +1 2,800installs nodejs-core A skill for editing and debugging Node.js core: nodejs/node commits and PRs, native crashes, V8, node-gyp, N-API and libuv. CX Codex +1 1,900installs documentation Structures technical documentation using the Diátaxis framework. It organizes content into tutorials, how-to guides, reference pages, and conceptual explanations. CX Codex +1 1,800installs octocat This skill handles Git and GitHub work through the gh CLI. Use it for pull requests, issues, CI checks, branch cleanup, and repository history. CX Codex +1 1,300installs init Builds AGENTS.md from what an agent cannot learn by reading the repo: odd commands, landmines, local conventions. Everything else gets cut. CX Codex +1 1,300installs oauth Implements OAuth 2.0 and 2.1 authorization flows in Fastify applications using PKCE and token rotation. Helps configure JWT validation hooks, session handling, and secure API routes. CX Codex +1 1,300installs linting-neostandard-eslint9 Configures ESLint v9 flat config and neostandard for JavaScript and TypeScript repositories. Helps migrate away from legacy .eslintrc files and set up CI checks. CX Codex +1 1,100installs snipgrapher Generates polished code snippet images in PNG, SVG, and WebP using the snipgrapher CLI. Useful for documentation, release notes, and social media posts. CX Codex +1 955installs