Skill Creator

Skill Creator

A skill for creating new skills and iteratively improving them.

At a high level, the process of creating a skill goes like this:

• Decide what you want the skill to do and roughly how it should do it
• Write a draft of the skill
• Create a few test prompts and run claude-with-access-to-the-skill on them
• Help the user evaluate the results both qualitatively and quantitatively
  • While the runs happen in the background, draft some quantitative evals if there aren't any (if there are some, you can either use as is or modify if you feel something needs to change about them). Then explain them to the user (or if they already existed, explain the ones that already exist)
  • Use the eval-viewer/generate_review.py script to show the user the results for them to look at, and also let them look at the quantitative metrics
• Rewrite the skill based on feedback from the user's evaluation of the results (and also if there are any glaring flaws that become apparent from the quantitative benchmarks)
• Repeat until you're satisfied
• Expand the test set and try again at larger scale

Your job when using this skill is to figure out where the user is in this process and then jump in and help them progress through these stages. So for instance, maybe they're like "I want to make a skill for X". You can help narrow down what they mean, write a draft, write the test cases, figure out how they want to evaluate, run all the prompts, and repeat.

On the other hand, maybe they already have a draft of the skill. In this case you can go straight to the eval/iterate part of the loop.

Of course, you should always be flexible and if the user is like "I don't need to run a bunch of evaluations, just vibe with me", you can do that instead.

Then after the skill is done (but again, the order is flexible), you can also run the skill description improver, which we have a whole separate script for, to optimize the triggering of the skill.

Cool? Cool.

Communicating with the user

The skill creator is liable to be used by people across a wide range of familiarity with coding jargon. If you haven't heard (and how could you, it's only very recently that it started), there's a trend now where the power of Claude is inspiring plumbers to open up their terminals, parents and grandparents to google "how to install npm". On the other hand, the bulk of users are probably fairly computer-literate.

So please pay attention to context cues to understand how to phrase your communication! In the default case, just to give you some idea:

• "evaluation" and "benchmark" are borderline, but OK
• for "JSON" and "assertion" you want to see serious cues from the user that they know what those things are before using them without explaining them

It's OK to briefly explain terms if you're in doubt, and feel free to clarify terms with a short definition if you're unsure if the user will get it.

Using AskUserQuestion (Critical — Read This)

Use the AskUserQuestion tool aggressively at every decision point. Do not ask open-ended text questions in conversation when structured choices exist. This is the single biggest UX improvement you can make — users juggle multiple windows and may not have looked at this conversation in 20 minutes.

Every AskUserQuestion MUST follow this structure:

1. Re-ground: State the skill name, current phase, and what just happened (1-2 sentences). The user may have context-switched away.
2. Simplify: Explain the decision in plain language. No function names or internal jargon. Say what it DOES, not what it's called.
3. Recommend: Lead with your recommendation and a one-line reason why. If options involve effort, show both scales: (human: ~X min / Claude: ~Y min).
4. Options: Provide 2-4 concrete, lettered choices. Each option should be a clear action, not an abstract concept.

Rules:

• One decision per question — never batch unrelated choices
• Provide an escape hatch ("Other" is always implicit in AskUserQuestion)
• Accept the user's choice — nudge on tradeoffs but never refuse to proceed
• Skip the question if there's an obvious answer with no tradeoffs (just state what you'll do)

Creating a skill

Capture Intent

Start by understanding the user's intent. The current conversation might already contain a workflow the user wants to capture (e.g., they say "turn this into a skill"). If so, extract answers from the conversation history first — the tools used, the sequence of steps, corrections the user made, input/output formats observed. The user may need to fill the gaps, and should confirm before proceeding to the next step.

1. What should this skill enable Claude to do?
2. When should this skill trigger? (what user phrases/contexts)
3. What's the expected output format?
4. Should we set up test cases to verify the skill works?

After extracting answers from conversation history (or asking questions 1-3), use AskUserQuestion to confirm the skill type and testing strategy:

Creating skill "[name]" — here's what I understand so far:
- Purpose: [1-sentence summary]
- Triggers on: [key phrases]
- Output: [format]

RECOMMENDATION: [Objective/Subjective/Hybrid] skill → [suggested testing approach]

Options:
A) Objective output (files, code, data) — set up automated test cases (Recommended if output is verifiable)
B) Subjective output (writing, design) — qualitative human review only
C) Hybrid — automated checks for structure, human review for quality
D) Skip testing for now — just build the skill and iterate by feel

This upfront classification drives the entire evaluation strategy downstream. Get it right here to avoid wasted effort later.

Prior Art Research (Do Not Skip)

The user's private methodology — their domain rules, workflow decisions, competitive edge — is what makes a skill valuable. No public repo can provide that. But the user shouldn't waste time reinventing infrastructure (API clients, auth flows, rate limiting) when mature tools exist. Prior art research finds building blocks for the infrastructure layer so the skill can focus on encoding the user's unique methodology.

Search these channels in order (use subagents for 4-8 in parallel):

Channels 1-3 surface the user's own proven patterns and existing integrations. Channels 4-8 find public infrastructure. The user's private SOP always takes precedence — public tools are building blocks, not replacements. In competitive domains (finance, trading, proprietary operations), the valuable methodology will never be public.

If a public MCP server or skill is found, clone it and verify — don't trust the README:

1. Read the actual source code — many projects have polished READMEs on hollow codebases
2. Verify auth method — does it match how the API actually authenticates? (X-Api-Key headers vs Bearer vs OAuth — many get this wrong)
3. Check test coverage — zero tests = prototype, not production-grade
4. Check maintenance — last commit date, open issue count, response to bug reports
5. Check environment compatibility — proxy/network assumptions, hardcoded DNS/IPs, region locks
6. Check license — MIT/Apache is fine; GPL/SSPL may conflict with proprietary use
7. Check dependency weight — huge dependency trees create conflict and security surface

Decision matrix:

After research completes, present findings via AskUserQuestion:

Research complete for "[skill-name]". Here's what I found:

[1-2 sentence summary of what exists publicly]

RECOMMENDATION: [ADOPT / EXTEND / BUILD] because [one-line reason]

Options:
A) Adopt [tool/MCP X] for infrastructure, build methodology layer on top (Recommended)
B) Extend [partial tool Y] — use what works, fill gaps in the skill
C) Build from scratch — nothing found matches well enough
D) Show me the detailed findings before I decide

When in doubt, bias toward adopting mature infrastructure for the plumbing layer and building custom logic for the methodology layer — that's where the value lives.

Interview and Research

Proactively ask questions about edge cases, input/output formats, example files, success criteria, and dependencies. Wait to write test prompts until you've got this part ironed out.

Check available MCPs - if useful for research (searching docs, finding similar skills, looking up best practices), research in parallel via subagents if available, otherwise inline. Come prepared with context to reduce burden on the user.

Write the SKILL.md

Based on the user interview, fill in these components:

• name: Skill identifier
• description: When to trigger, what it does. This is the primary triggering mechanism - include both what the skill does AND specific contexts for when to use it. All "when to use" info goes here, not in the body. Note: currently Claude has a tendency to "undertrigger" skills -- to not use them when they'd be useful. To combat this, please make the skill descriptions a little bit "pushy". So for instance, instead of "How to build a simple fast dashboard to display internal Anthropic data.", you might write "How to build a simple fast dashboard to display internal Anthropic data. Make sure to use this skill whenever the user mentions dashboards, data visualization, internal metrics, or wants to display any kind of company data, even if they don't explicitly ask for a 'dashboard.'"
• compatibility: Required tools, dependencies (optional, rarely needed)
• the rest of the skill :)

Skill Writing Guide

Anatomy of a Skill

skill-name/
├── SKILL.md (required)
│   ├── YAML frontmatter (name, description required)
│   └── Markdown instructions
└── Bundled Resources (optional)
    ├── scripts/    - Executable code for deterministic/repetitive tasks
    ├── references/ - Docs loaded into context as needed
    └── assets/     - Files used in output (templates, icons, fonts)

YAML Frontmatter Reference

All frontmatter fields except description are optional. Configure skill behavior using these fields between --- markers:

---
name: my-skill
description: What this skill does and when to use it. Use when...
context: fork
agent: general-purpose
argument-hint: [topic]
---

Special placeholder: $ARGUMENTS in skill content is replaced with text the user provides after the skill name. For example, /deep-research quantum computing replaces $ARGUMENTS with quantum computing.

Inline vs Fork: Critical Decision

This is the most important architectural decision when designing a skill. Choosing wrong will silently break your skill's core capabilities.

CRITICAL CONSTRAINT: Subagents cannot spawn other subagents. A skill running with context: fork (as a subagent) CANNOT:

• Use the Task tool to spawn parallel exploration agents
• Use the Skill tool to invoke other skills
• Orchestrate any multi-agent workflow

Decision guide:

Example: Orchestrator skill (MUST be inline):

---
name: product-analysis
description: Multi-path parallel product analysis with cross-model synthesis
---

# Orchestrates parallel agents — inline is REQUIRED
1. Auto-detect available tools (which codex, etc.)
2. Launch 3-5 Task agents in parallel (Explore subagents)
3. Optionally invoke /competitors-analysis via Skill tool
4. Synthesize all results

Example: Specialist skill (fork is correct):

---
name: deep-research
description: Research a topic thoroughly using multiple sources
context: fork
agent: Explore
---

Research $ARGUMENTS thoroughly:
1. Find relevant files using Glob and Grep
2. Read and analyze the code
3. Summarize findings with specific file references

Example: Reference skill (inline, no task):

---
name: api-conventions
description: API design patterns for this codebase
---

When writing API endpoints:
- Use RESTful naming conventions
- Return consistent error formats

Composable Skill Design (Orthogonality)

Skills should be orthogonal: each skill handles one concern, and they combine through composition.

Pattern: Orchestrator (inline) calls Specialist (fork)

product-analysis (inline, orchestrator)
  ├─ Task agents for parallel exploration
  ├─ Skill('competitors-analysis', 'X') → fork subagent
  └─ Synthesizes all results

competitors-analysis (fork, specialist)
  └─ Single focused task: analyze one competitor codebase

Rules for composability:

1. The caller must be inline (no context: fork) to use Task/Skill tools
2. The callee should use context: fork to run in isolated subagent context
3. Each skill has a single responsibility — don't mix orchestration with execution
4. Share methodology via references (e.g., checklists, templates), not by duplicating code

Auto-Detection Over Manual Flags

Never add manual flags for capabilities that can be auto-detected. Instead of requiring users to pass --with-codex or --verbose, detect capabilities at runtime:

# Good: Auto-detect and inform
Step 0: Check available tools
  - `which codex` → If found, inform user and enable cross-model analysis
  - `ls package.json` → If found, tailor prompts for Node.js project
  - `which docker` → If found, enable container-based execution

# Bad: Manual flags
argument-hint: [scope] [--with-codex] [--docker] [--verbose]

Principle: Capabilities auto-detect, user decides scope. A skill should discover what it CAN do and act accordingly, not require users to remember what tools are installed.

Invocation Control

Progressive Disclosure

Skills use a three-level loading system:

1. Metadata (name + description) - Always in context (~100 words)
2. SKILL.md body - In context whenever skill triggers (<500 lines ideal)
3. Bundled resources - As needed (unlimited, scripts can execute without loading)

These word counts are approximate and you can feel free to go longer if needed.

Key patterns:

• Keep SKILL.md under 500 lines; if you're approaching this limit, add an additional layer of hierarchy along with clear pointers about where the model using the skill should go next to follow up.
• Reference files clearly from SKILL.md with guidance on when to read them
• For large reference files (>300 lines), include a table of contents

Domain organization: When a skill supports multiple domains/frameworks, organize by variant:

cloud-deploy/
├── SKILL.md (workflow + selection)
└── references/
    ├── aws.md
    ├── gcp.md
    └── azure.md

Claude reads only the relevant reference file.

Principle of Lack of Surprise

This goes without saying, but skills must not contain malware, exploit code, or any content that could compromise system security. A skill's contents should not surprise the user in their intent if described. Don't go along with requests to create misleading skills or skills designed to facilitate unauthorized access, data exfiltration, or other malicious activities. Things like a "roleplay as an XYZ" are OK though.

Writing Patterns

Prefer using the imperative form in instructions.

Defining output formats - You can do it like this:

## Report structure
ALWAYS use this exact template:
# [Title]
## Executive summary
## Key findings
## Recommendations

Examples pattern - It's useful to include examples. You can format them like this (but if "Input" and "Output" are in the examples you might want to deviate a little):

## Commit message format
**Example 1:**
Input: Added user authentication with JWT tokens
Output: feat(auth): implement JWT-based authentication

Writing Style

Try to explain to the model why things are important in lieu of heavy-handed musty MUSTs. Use theory of mind and try to make the skill general and not super-narrow to specific examples. Start by writing a draft and then look at it with fresh eyes and improve it.

Bundled Resources

Scripts (scripts/)

Executable code (Python/Bash/etc.) for tasks that require deterministic reliability or are repeatedly rewritten.

• When to include: When the same code is being rewritten repeatedly or deterministic reliability is needed
• Example: scripts/rotate_pdf.py for PDF rotation tasks
• Benefits: Token efficient, deterministic, may be executed without loading into context
• Note: Scripts may still need to be read by Claude for patching or environment-specific adjustments

References (references/)

Documentation and reference material intended to be loaded as needed into context to inform Claude's process and thinking.

• When to include: For documentation that Claude should reference while working
• Examples: references/finance.md for financial schemas, references/mnda.md for company NDA template
• Use cases: Database schemas, API documentation, domain knowledge, company policies, detailed workflow guides
• Benefits: Keeps SKILL.md lean, loaded only when Claude determines it's needed
• Best practice: If files are large (>10k words), include grep search patterns in SKILL.md
• Avoid duplication: Information should live in either SKILL.md or references files, not both

Assets (assets/)

Files not intended to be loaded into context, but rather used within the output Claude produces.

• When to include: When the skill needs files that will be used in the final output
• Examples: assets/logo.png for brand assets, assets/slides.pptx for PowerPoint templates
• Use cases: Templates, images, icons, boilerplate code, fonts, sample documents

Privacy and Path References

CRITICAL: Skills intended for public distribution must not contain user-specific or company-specific information:

• Forbidden: Absolute paths to user directories (/home/username/, /Users/username/)
• Forbidden: Personal usernames, company names, product names
• Forbidden: Hardcoded skill installation paths like ~/.claude/skills/
• Allowed: Relative paths within the skill bundle (scripts/example.py, references/guide.md)
• Allowed: Standard placeholders (~/workspace/project, username, your-company)

Versioning

CRITICAL: Skills should NOT contain version history or version numbers in SKILL.md:

• Forbidden: Version sections (## Version, ## Changelog) in SKILL.md
• Correct location: Skill versions are tracked in marketplace.json under plugins[].version
• Rationale: Marketplace infrastructure manages versioning; SKILL.md should be timeless content

Reference File Naming

Filenames must be self-explanatory without reading contents.

Pattern: <content-type>_<specificity>.md

Examples:

• Bad: commands.md, cli_usage.md, reference.md
• Good: script_parameters.md, api_endpoints.md, database_schema.md

Test: Can someone understand the file's contents from the name alone?

Skill Creation Best Practice

Anthropic has wrote skill authoring best practices, you SHOULD retrieve it before you create or update any skills, the link is https://platform.claude.com/docs/en/agents-and-tools/agent-skills/best-practices.md

Test Cases

After writing the skill draft, come up with 2-3 realistic test prompts — the kind of thing a real user would actually say. Present them via AskUserQuestion:

Skill draft is ready. Here are [N] test cases I'd like to run:

1. "[test prompt 1]" — tests [what aspect]
2. "[test prompt 2]" — tests [what aspect]
3. "[test prompt 3]" — tests [what aspect]

Each test runs the skill + a baseline (no skill) for comparison.
Estimated time: ~[X] minutes total.

RECOMMENDATION: Run all [N] test cases now.

Options:
A) Run all test cases (Recommended)
B) Run test cases, but let me modify them first
C) Add more test cases before running
D) Skip testing — the skill looks good enough to ship

Save test cases to evals/evals.json. Don't write assertions yet — just the prompts. You'll draft assertions in the next step while the runs are in progress.

{
  "skill_name": "example-skill",
  "evals": [
    {
      "id": 1,
      "prompt": "User's task prompt",
      "expected_output": "Description of expected result",
      "files": []
    }
  ]
}

See references/schemas.md for the full schema (including the assertions field, which you'll add later).

Running and evaluating test cases

This section is one continuous sequence — don't stop partway through. Do NOT use /skill-test or any other testing skill.

Put results in <skill-name>-workspace/ as a sibling to the skill directory. Within the workspace, organize results by iteration (iteration-1/, iteration-2/, etc.) and within that, each test case gets a directory (eval-0/, eval-1/, etc.). Don't create all of this upfront — just create directories as you go.

Step 1: Spawn all runs (with-skill AND baseline) in the same turn

For each test case, spawn two subagents in the same turn — one with the skill, one without. This is important: don't spawn the with-skill runs first and then come back for baselines later. Launch everything at once so it all finishes around the same time.

With-skill run:

Execute this task:
- Skill path: <path-to-skill>
- Task: <eval prompt>
- Input files: <eval files if any, or "none">
- Save outputs to: <workspace>/iteration-<N>/eval-<ID>/with_skill/outputs/
- Outputs to save: <what the user cares about — e.g., "the .docx file", "the final CSV">

Baseline run (same prompt, but the baseline depends on context):

• Creating a new skill: no skill at all. Same prompt, no skill path, save to without_skill/outputs/.
• Improving an existing skill: the old version. Before editing, snapshot the skill (cp -r <skill-path> <workspace>/skill-snapshot/), then point the baseline subagent at the snapshot. Save to old_skill/outputs/.

Write an eval_metadata.json for each test case (assertions can be empty for now). Give each eval a descriptive name based on what it's testing — not just "eval-0". Use this name for the directory too. If this iteration uses new or modified eval prompts, create these files for each new eval directory — don't assume they carry over from previous iterations.

{
  "eval_id": 0,
  "eval_name": "descriptive-name-here",
  "prompt": "The user's task prompt",
  "assertions": []
}

Step 2: While runs are in progress, draft assertions

Don't just wait for the runs to finish — you can use this time productively. Draft quantitative assertions for each test case and explain them to the user. If assertions already exist in evals/evals.json, review them and explain what they check.

Good assertions are objectively verifiable and have descriptive names — they should read clearly in the benchmark viewer so someone glancing at the results immediately understands what each one checks. Subjective skills (writing style, design quality) are better evaluated qualitatively — don't force assertions onto things that need human judgment.

Update the eval_metadata.json files and evals/evals.json with the assertions once drafted. Also explain to the user what they'll see in the viewer — both the qualitative outputs and the quantitative benchmark.

Step 3: As runs complete, capture timing data

When each subagent task completes, you receive a notification containing total_tokens and duration_ms. Save this data immediately to timing.json in the run directory:

{
  "total_tokens": 84852,
  "duration_ms": 23332,
  "total_duration_seconds": 23.3
}

This is the only opportunity to capture this data — it comes through the task notification and isn't persisted elsewhere. Process each notification as it arrives rather than trying to batch them.

Step 4: Grade, aggregate, and launch the viewer

Once all runs are done:

1. Grade each run — spawn a grader subagent (or grade inline) that reads agents/grader.md and evaluates each assertion against the outputs. Save results to grading.json in each run directory. The grading.json expectations array must use the fields text, passed, and evidence (not name/met/details or other variants) — the viewer depends on these exact field names. For assertions that can be checked programmatically, write and run a script rather than eyeballing it — scripts are faster, more reliable, and can be reused across iterations.

2. Aggregate into benchmark — run the aggregation script from the skill-creator directory:

   python -m scripts.aggregate_benchmark <workspace>/iteration-N --skill-name <name>
   

   This produces benchmark.json and benchmark.md with pass_rate, time, and tokens for each configuration, with mean +/- stddev and the delta. If generating benchmark.json manually, see references/schemas.md for the exact schema the viewer expects. Put each with_skill version before its baseline counterpart.

3. Do an analyst pass — read the benchmark data and surface patterns the aggregate stats might hide. See agents/analyzer.md (the "Analyzing Benchmark Results" section) for what to look for — things like assertions that always pass regardless of skill (non-discriminating), high-variance evals (possibly flaky), and time/token tradeoffs.

4. Launch the viewer with both qualitative outputs and quantitative data:

   nohup python <skill-creator-path>/eval-viewer/generate_review.py \
     <workspace>/iteration-N \
     --skill-name "my-skill" \
     --benchmark <workspace>/iteration-N/benchmark.json \
     > /dev/null 2>&1 &
   VIEWER_PID=$!
   

   For iteration 2+, also pass --previous-workspace <workspace>/iteration-<N-1>.

   Cowork / headless environments: If webbrowser.open() is not available or the environment has no display, use --static <output_path> to write a standalone HTML file instead of starting a server. Feedback will be downloaded as a feedback.json file when the user clicks "Submit All Reviews". After download, copy feedback.json into the workspace directory for the next iteration to pick up.

Note: please use generate_review.py to create the viewer; there's no need to write custom HTML.

5. Tell the user via AskUserQuestion:

Results are ready! I've opened the eval viewer in your browser.

- "Outputs" tab: click through each test case, leave feedback in the textbox
- "Benchmark" tab: quantitative comparison (pass rates, timing, tokens)

Take your time reviewing. When you're done, come back here.

Options:
A) I've finished reviewing — read my feedback and improve the skill
B) I have questions about the results before giving feedback
C) Results look good enough — skip iteration, let's package the skill
D) Results need major rework — let's discuss before iterating

What the user sees in the viewer

The "Outputs" tab shows one test case at a time:

• Prompt: the task that was given
• Output: the files the skill produced, rendered inline where possible
• Previous Output (iteration 2+): collapsed section showing last iteration's output
• Formal Grades (if grading was run): collapsed section showing assertion pass/fail
• Feedback: a textbox that auto-saves as they type
• Previous Feedback (iteration 2+): their comments from last time, shown below the textbox

The "Benchmark" tab shows the stats summary: pass rates, timing, and token usage for each configuration, with per-eval breakdowns and analyst observations.

Navigation is via prev/next buttons or arrow keys. When done, they click "Submit All Reviews" which saves all feedback to feedback.json.

Step 5: Read the feedback

When the user tells you they're done, read feedback.json:

{
  "reviews": [
    {"run_id": "eval-0-with_skill", "feedback": "the chart is missing axis labels", "timestamp": "..."},
    {"run_id": "eval-1-with_skill", "feedback": "", "timestamp": "..."},
    {"run_id": "eval-2-with_skill", "feedback": "perfect, love this", "timestamp": "..."}
  ],
  "status": "complete"
}

Empty feedback means the user thought it was fine. Focus your improvements on the test cases where the user had specific complaints.

Kill the viewer server when you're done with it:

kill $VIEWER_PID 2>/dev/null

Improving the skill

This is the heart of the loop. You've run the test cases, the user has reviewed the results, and now you need to make the skill better based on their feedback.

How to think about improvements

1. Generalize from the feedback. The big picture thing that's happening here is that we're trying to create skills that can be used a million times (maybe literally, maybe even more who knows) across many different prompts. Here you and the user are iterating on only a few examples over and over again because it helps move faster. The user knows these examples in and out and it's quick for them to assess new outputs. But if the skill you and the user are codeveloping works only for those examples, it's useless. Rather than put in fiddly overfitty changes, or oppressively constrictive MUSTs, if there's some stubborn issue, you might try branching out and using different metaphors, or recommending different patterns of working. It's relatively cheap to try and maybe you'll land on something great.

2. Keep the prompt lean. Remove things that aren't pulling their weight. Make sure to read the transcripts, not just the final outputs — if it looks like the skill is making the model waste a bunch of time doing things that are unproductive, you can try getting rid of the parts of the skill that are making it do that and seeing what happens.

3. Explain the why. Try hard to explain the why behind everything you're asking the model to do. Today's LLMs are smart. They have good theory of mind and when given a good harness can go beyond rote instructions and really make things happen. Even if the feedback from the user is terse or frustrated, try to actually understand the task and why the user is writing what they wrote, and what they actually wrote, and then transmit this understanding into the instructions. If you find yourself writing ALWAYS or NEVER in all caps, or using super rigid structures, that's a yellow flag — if possible, reframe and explain the reasoning so that the model understands why the thing you're asking for is important. That's a more humane, powerful, and effective approach.

4. Look for repeated work across test cases. Read the transcripts from the test runs and notice if the subagents all independently wrote similar helper scripts or took the same multi-step approach to something. If all 3 test cases resulted in the subagent writing a create_docx.py or a build_chart.py, that's a strong signal the skill should bundle that script. Write it once, put it in scripts/, and tell the skill to use it. This saves every future invocation from reinventing the wheel.

This task is pretty important (we are trying to create billions a year in economic value here!) and your thinking time is not the blocker; take your time and really mull things over. I'd suggest writing a draft revision and then looking at it anew and making improvements. Really do your best to get into the head of the user and understand what they want and need.

After analyzing feedback, present your improvement plan via AskUserQuestion:

I've read the feedback from [N] test cases. [X] had specific complaints, [Y] looked good.

Key issues:
- [Issue 1]: [plain-language summary]
- [Issue 2]: [plain-language summary]

RECOMMENDATION: [strategy] because [reason]

Options:
A) Iterative refinement — targeted fixes for the specific issues above (Recommended)
B) Structural redesign — the core approach needs rethinking
C) Bundle a script — I noticed all test runs independently wrote similar code for [X]
D) Expand test set first — add [N] more test cases to avoid overfitting to these examples

The iteration loop

After improving the skill:

1. Apply your improvements to the skill
2. Rerun all test cases into a new iteration-<N+1>/ directory, including baseline runs. If you're creating a new skill, the baseline is always without_skill (no skill) — that stays the same across iterations. If you're improving an existing skill, use your judgment on what makes sense as the baseline: the original version the user came in with, or the previous iteration.
3. Launch the reviewer with --previous-workspace pointing at the previous iteration
4. Wait for the user to review and tell you they're done
5. Read the new feedback, improve again, repeat

At the end of each iteration, use AskUserQuestion as a checkpoint:

Iteration [N] complete. Results: [pass_rate]% assertions passing, [delta vs previous].

Options:
A) Continue iterating — I see more room for improvement
B) Accept this version — it's good enough, let's move to packaging
C) Revert to previous iteration — this round made things worse
D) Run blind comparison — rigorously compare this version vs the previous one

Keep going until:

• The user says they're happy
• The feedback is all empty (everything looks good)
• You're not making meaningful progress

Advanced: Blind comparison

For situations where you want a more rigorous comparison between two versions of a skill (e.g., the user asks "is the new version actually better?"), there's a blind comparison system. Read agents/comparator.md and agents/analyzer.md for the details. The basic idea is: give two outputs to an independent agent without telling it which is which, and let it judge quality. Then analyze why the winner won.

This is optional, requires subagents, and most users won't need it. The human review loop is usually sufficient.

Description Optimization

The description field in SKILL.md frontmatter is the primary mechanism that determines whether Claude invokes a skill. After creating or improving a skill, offer to optimize the description for better triggering accuracy.

Step 1: Generate trigger eval queries

Create 20 eval queries — a mix of should-trigger and should-not-trigger. Save as JSON:

[
  {"query": "the user prompt", "should_trigger": true},
  {"query": "another prompt", "should_trigger": false}
]

The queries must be realistic and something a Claude Code or Claude.ai user would actually type. Not abstract requests, but requests that are concrete and specific and have a good amount of detail. For instance, file paths, personal context about the user's job or situation, column names and values, company names, URLs. A little bit of backstory. Some might be in lowercase or contain abbreviations or typos or casual speech. Use a mix of different lengths, and focus on edge cases rather than making them clear-cut (the user will get a chance to sign off on them).

Bad: "Format this data", "Extract text from PDF", "Create a chart"

Good: "ok so my boss just sent me this xlsx file (its in my downloads, called something like 'Q4 sales final FINAL v2.xlsx') and she wants me to add a column that shows the profit margin as a percentage. The revenue is in column C and costs are in column D i think"

For the *should-trigger

…(truncated)