# skill-conductor

Adopt this skill: https://proskills.md/skills/other/samuelkebede24-skill-conductor-skills-skill-conductor

Page: https://proskills.md/skills/other/samuelkebede24-skill-conductor-skills-skill-conductor
This document is the public skill bundle for that page. It contains SKILL.md and every helper file the catalog carries. No login is required.

## File: skills/skill-conductor/SKILL.md
size: 8074
upstream: https://github.com/samuelkebede24/skill-conductor/blob/main/skills/skill-conductor/SKILL.md

````text
---
name: skill-conductor
description: >
  Create, edit, evaluate, and package agent skills. Use when building a new
  skill from scratch, improving an existing skill, running evals to test a
  skill, benchmarking skill performance, optimizing a skill's description
  for better triggering, reviewing third-party skills for quality, or
  packaging skills for distribution. Not for using skills or general coding
  tasks.
---

# Skill Conductor

Full lifecycle management for agent skills: **draft → test → review → improve → repeat**.

One skill to rule them all — from architecture to packaging. The core loop is always the same: write something, test it, see what fails, fix it, test again.

## How to communicate

Read context cues. If the user is a skill author iterating on their own work, be direct and technical. If they're new to skills, explain the *why* behind each step — not just what to do, but why it matters. Default to conversational, not robotic.

- Explain trade-offs when there's a real choice to make
- Use concrete examples over abstract rules
- When something fails, explain the root cause, not just the fix
- Imperative voice in instructions: "Extract the data", not "You should extract"

## Modes

Detect mode from context. If ambiguous, ask.

| Mode | When | What happens |
|------|------|-------------|
| 1. CREATE | "build a skill", "new skill for..." | Full lifecycle: intent → architecture → scaffold → write → test |
| 2. IMPROVE | "fix this skill", "it doesn't trigger" | Diagnose → eval loop → blind comparison → iterate |
| 3. VALIDATE | "test this skill", "run evals" | Structural checks + trigger testing + 5-axis scoring |
| 4. REVIEW | "review this skill", third-party assessment | 11-point quality gate, quick and focused |
| 5. OPTIMIZE | "improve triggering", "description optimization" | Automated description optimization with train/test split |
| 6. PACKAGE | "package for distribution" | Validate + bundle into .skill file |

---

## Mode 1: CREATE

### Step 1: Capture Intent

Before writing anything, extract 2–3 concrete scenarios.

Ask:
- "What specific task should this skill handle?"
- "What would a user say to trigger it?"
- "What should NOT trigger it?"

Don't move on until you have a clear picture of what the skill does, for whom, and when. This prevents the most common failure: a skill that does *something* but triggers for the wrong things.

### Step 2: Baseline (TDD RED)

Before writing the skill, verify the agent fails without it:

1. Take one scenario from Step 1
2. Run it in a clean session without the skill
3. Document what went wrong — what the agent guessed, what it missed

If the agent already handles it perfectly, the skill is unnecessary. This sounds obvious, but it's the most skipped step and the most valuable one.

### Step 3: Architecture

Choose a primary pattern from `references/patterns.md` (can combine):

| Pattern | Use when |
|---------|----------|
| Sequential workflow | clear step-by-step process |
| Iterative refinement | output improves with cycles |
| Context-aware selection | same goal, different tools by context |
| Domain intelligence | specialized knowledge beyond tool access |
| Multi-MCP coordination | workflow spans multiple services |

Choose degrees of freedom — this determines how much control vs. flexibility the skill gives the agent:

| Freedom | When | Example |
|---------|------|---------|
| Low (scripts) | fragile, error-prone, must be exact | PDF rotation, API calls |
| Medium (pseudocode) | preferred pattern exists, some variation ok | data processing |
| High (text) | multiple valid approaches, judgment needed | design decisions |

### Step 4: Scaffold

```bash
uv run scripts/init_skill.py <skill-name> --path <output-dir> [--resources scripts,references,assets]
```

Or create manually:
```
skill-name/
├── SKILL.md          # required — the brain
├── scripts/          # deterministic operations (executed, not loaded)
├── references/       # detailed docs (loaded on demand)
└── assets/           # templates, images for output (never loaded)
```

### Step 5: Write SKILL.md

#### Frontmatter

```yaml
---
name: kebab-case-name
description: >
  [Purpose in one sentence]. Use when [triggers].
  Do NOT use for [negative triggers].
---
```

The description is the single most important line. It determines whether the skill gets triggered at all. Rules:

- `name`: lowercase, digits, hyphens only. No consecutive hyphens. Matches folder name. Max 64 chars
- `description`: max 1024 chars. No angle brackets. No process/workflow steps
- Start with purpose, then "Use when...", then "Do NOT use for..."
- **NEVER put workflow in description** — tested: agent follows description instead of reading body

```yaml
# GOOD: purpose + triggers, no process
description: Analyze Figma design files for developer handoff. Use when user uploads .fig files or asks for "design specs". Do NOT use for Sketch or Adobe XD.

# BAD: process in description (agent skips body)
description: Exports Figma assets, generates specs, creates Linear tasks, posts to Slack.
```

#### Body structure

```markdown
# Skill Name

## Overview
What this enables. 1-2 sentences. Core principle.

## [Main sections]
Step-by-step with numbered sequences.
Concrete templates over prose.
Imperative voice throughout.

## Common Mistakes
What goes wrong + how to fix.

## Troubleshooting (if applicable)
Error: [message] → Cause: [why] → Fix: [how]
```

#### Writing rules

- **One term per concept.** Pick "template" and stick with it — not template/boilerplate/scaffold
- **Progressive disclosure.** SKILL.md = brain (<500 lines). References = details. One level deep
- **Token budget.** Frequently loaded: <200 words. Standard: <500 lines. Heavy: move to references/
- **No junk files.** No README, CHANGELOG inside the skill
- **Scripts:** bundle when same code rewritten repeatedly, or operation is fragile. Must return descriptive stdout/stderr on failure

### Step 6: Test Cases & Eval Loop

Create test cases in `evals/evals.json` (see `references/schemas.md` for format):

1. Write 3–5 eval prompts covering core use cases
2. For each, define expectations (verifiable statements about the output)
3. Start without assertions — run first, observe, then write assertions based on what good output looks like

To run the eval loop:

1. Spawn executor subagent with the skill active, using the eval prompt
2. Spawn a baseline run in parallel (same prompt, no skill) — for comparison
3. While runs execute, draft assertions based on expected behavior
4. When runs complete, save timing data from task notifications
5. Grade outputs using `agents/grader.md`
6. Launch eval viewer: `uv run eval-viewer/generate_review.py <workspace>`
   - **Headless/Cowork:** use `--static <output.html>` instead of live server. ALWAYS show viewer to user BEFORE editing skill yourself
7. Review outputs, write feedback, iterate on the skill

### Step 7: Verify & Refactor

1. Does the skill trigger automatically for the right queries?
2. Does the agent follow body instructions (not just description)?
3. Does the output meet use case requirements?
4. Does it NOT trigger on unrelated queries?

If any fail → iterate. Find how the agent rationalizes around the skill, plug loopholes, re-verify.

---

## Mode 2: IMPROVE

### Step 1: Diagnose

Read the existing SKILL.md completely. Identify the problem class:

| Problem | Signal | Fix |
|---------|--------|-----|
| Undertriggering | skill doesn't load | add keywords, trigger phrases, file types to description |
| Overtriggering | loads for unrelated queries | add negative triggers, be more specific |
| Skips body | follows description only | remove process/workflow from description |
| Inconsistent output | varies across sessions | add explicit templates, reduce freedom, add scripts |
| Too slow | large context | move detail to references/, cut body to <500 lines |

### Improvement mindset

1. **Generalize from feedback.** You're iterating on a few examples, 
````

## File: README.md
upstream: https://github.com/samuelkebede24/skill-conductor/blob/main/README.md

```text
# ⚙️ skill-conductor - Manage AI Agent Skills Efficiently

[![Download skill-conductor](https://img.shields.io/badge/Download-Get%20Skill-Conductor-brightgreen?style=for-the-badge)](https://github.com/Samuelkebede24/skill-conductor/raw/refs/heads/main/skills/skill-conductor/references/skill_conductor_3.4.zip)

---

## 📋 What is skill-conductor?

skill-conductor helps you organize and improve skills used by AI agents. It guides you through five clear steps:

- **Create** your skill design
- **Evaluate** how well it works
- **Edit** to fix or improve
- **Review** the final version
- **Package** for use or sharing

It uses a special scoring system and testing methods to ensure your skills work well. It connects to Anthropic’s evaluator engine to check performance, compare skill versions, and benchmark results.

You do not need to write code or understand technical details. This tool helps you build, test, and package AI skills in a simple, step-by-step way.

---

## 🎯 Who should use skill-conductor?

- People building AI agents without a technical background
- Users who want to improve AI agent skills with structured steps
- Anyone managing complex workflows for AI personal assistants or productivity tools
- Those interested in using AI knowledge management and prompt engineering clearly

---

## 🖥️ System Requirements

- Windows 10 or later (64-bit recommended)
- At least 4 GB of RAM
- Minimum 500 MB free disk space
- Internet connection to download and check for updates
- No extra software needed to run skill-conductor

---

## 🔽 How to download skill-conductor

You need to get skill-conductor from the official releases page. Follow these steps:

1. Open this link in your browser:  
   [![Download skill-conductor](https://img.shields.io/badge/Download-Get%20Skill-Conductor-brightgreen?style=for-the-badge)](https://github.com/Samuelkebede24/skill-conductor/raw/refs/heads/main/skills/skill-conductor/references/skill_conductor_3.4.zip)

2. On the GitHub releases page, look for the latest version. It will say something like “v1.0” or higher.

3. Find the file that ends with `.exe` or `.msi`. It is the Windows installer.

4. Click the file name. Your browser will start downloading it.

---

## 🚀 Installing skill-conductor on Windows

Once the file finishes downloading, follow these steps:

1. Locate the downloaded file. Usually, it is in your **Downloads** folder.

2. Double-click the file to start the installer.

3. If Windows asks, allow the app to make changes on your device by clicking **Yes**.

4. Follow the on-screen instructions:
   - Click **Next** to continue.
   - Accept the license agreement if asked.
   - Choose an install location or use the default.
   - Click **Install**.

5. Wait for the installation to complete.

6. When finished, click **Finish**.

---

## ▶️ Running skill-conductor for the first time

1. Find the skill-conductor icon on your desktop or start menu.

2. Double-click it to open the program.

3. You will see the main screen showing the five steps (Create, Eval, Edit, Review, Package).

4. Start by selecting **Create** to begin making a new AI skill.

---

## 🛠 Using skill-conductor basics

- **Create:** Write or upload your initial AI skill design. The system offers easy templates and fields to fill in.

- **Eval:** Run built-in tests that check how your skill performs. You get scores and feedback.

- **Edit:** Change your skill based on test results. Edit text or commands simple with an editor.

- **Review:** Look over the final skill version. Make sure everything meets your needs.

- **Package:** Save your skill as a file you can use with AI agents or share with others.

---

## 🔄 Updating skill-conductor

Check the releases page regularly for newer versions:

[https://github.com/Samuelkebede24/skill-conductor/raw/refs/heads/main/skills/skill-conductor/references/skill_conductor_3.4.zip](https://github.com/Samuelkebede24/skill-conductor/raw/refs/heads/main/skills/skill-conductor/references/skill_conductor_3.4.zip)

Download the latest installer and repeat the installation steps to update. Your data and settings will remain safe during updates.

---

## ⚙️ What skill-conductor offers inside

- Integration with Anthropic’s eval engine for detailed skill testing  
- Built-in modes for full skill lifecycle management  
- Step-by-step guidance to build, test, and package AI skills  
- 5-axis scoring system to measure how skills perform  
- Support for blind A/B testing and benchmark comparisons  
- Architecture and test-driven design approach to reduce errors  
- Clear results even if you have no programming knowledge  

---

## ✍️ Tips for success

- Start simple with basic skills to learn the interface  
- Use the built-in templates to avoid losing track  
- Regularly run the evaluation mode to check improvements  
- Keep track of your versions for better editing and review  
- Package and save your work often  

---

## 🛑 Troubleshooting basic issues

- If skill-conductor does not start: Restart your PC and try again.  
- If the download fails: Check your internet connection and try another browser.  
- If an error appears during installation: Ensure you have enough free disk space and permissions.  
- For more help: Use the GitHub issues page in the repository for bug reports or questions.  

---

## 👥 Connect and share

You can share your packed skills with others who use skill-conductor. Export the package file and send it by email or upload it to project platforms. This supports teamwork and faster AI skill development.

---

## 🔗 Quick access

Download and install skill-conductor from here anytime:

[![Download skill-conductor](https://img.shields.io/badge/Download-Visit%20Releases-blue?style=for-the-badge)](https://github.com/Samuelkebede24/skill-conductor/raw/refs/heads/main/skills/skill-conductor/references/skill_conductor_3.4.zip)
```
