skill-creator

name	skill-creator
description	Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

Skill Creator

A skill for creating new skills and iteratively improving them in Pi. Adapted from the upstream anthropics/skills repo for Pi's Agent Skills implementation.

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 them with the skill loaded
• 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.

Pi-Specific Notes

Pi skills follow the Agent Skills standard and use YAML frontmatter with required name and description fields. Pi warns about most specification violations but is lenient; unlike the standard, Pi does not require the skill name to match its parent directory.

Skill locations. Pi discovers global skills from $PI_CODING_AGENT_DIR/skills/ (configured in these dotfiles as ~/.config/pi/skills/) and ~/.agents/skills/; project skills from .pi/skills/ and .agents/skills/ in the current directory and ancestors; package skills from skills/ directories or pi.skills package entries; settings paths from the skills array; and explicit CLI paths via --skill <path>. In .pi/skills/ and $PI_CODING_AGENT_DIR/skills/, root .md files can be standalone skills. Directories containing SKILL.md are discovered recursively.

Skill loading. At startup Pi puts each skill's name and description into the system prompt. When a task matches, the model should use read to load the full SKILL.md. Users can force a skill with /skill:<name>, and extra text after the command is appended as User: <args>.

Eval runs use separate pi processes. Pi keeps the core minimal and does not ship built-in agent fanout, MCP, or a TodoList. For independent eval runs, launch separate pi --mode json --no-session ... processes from bash, a small Python orchestrator, or tmux. Use --no-skills --skill <path> to load exactly the skill under test, and use /skill:<name> ... in the prompt when you want to force the skill for output-quality benchmarks. Use --no-skills without --skill for a no-skill baseline. Keep any isolation flags consistent across configurations, and omit them only when the skill intentionally depends on project context, prompt templates, or extensions.

pi --mode json --no-session --no-context-files --no-prompt-templates \
  --no-skills --skill /path/to/skill \
  "/skill:my-skill Execute this eval task..." > transcript.jsonl

pi --mode json --no-session --no-context-files --no-prompt-templates \
  --no-skills "Execute the same eval task without the skill..." > baseline.jsonl

Opening files in browser. On macOS, use open /path/to/file.html to open HTML files. Use --static mode with generate_review.py to write standalone HTML, then open it.

Description optimization scripts use Pi. run_eval.py tests automatic skill triggering with pi --mode json; improve_description.py and run_loop.py use pi -p. They require the Pi CLI to be on PATH and authenticated. If the scripts are too heavy for the situation, iterate manually: draft trigger eval queries, review them with the user, and adjust the description by hand.

Communicating with the user

The skill creator is liable to be used by people across a wide range of familiarity with coding jargon. AI coding assistants are 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.

---

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 the model 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? Skills with objectively verifiable outputs (file transforms, data extraction, code generation, fixed workflow steps) benefit from test cases. Skills with subjective outputs (writing style, art) often don't need them. Suggest the appropriate default based on the skill type, but let the user decide.

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 Pi resources: project docs, existing skills, installed packages/extensions, scripts, and reference files. Pi has no MCP by default; if a useful extension/tool is available, use it, otherwise research inline with the built-in tools. For worthwhile parallel research, use separate Pi CLI runs or tmux; otherwise keep it inline to avoid orchestration overhead.

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: models tend to "undertrigger" skills -- to not use them when they'd be useful. To combat this, make the skill descriptions a little bit "pushy". So for instance, instead of "How to build a simple fast dashboard to display internal data.", you might write "How to build a simple fast dashboard to display internal 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)
• disable-model-invocation: Optional Pi field; set to true only for skills that should be hidden from automatic triggering and used via /skill:<name>
• 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)

Pi also supports root .md skills in .pi/skills/ and $PI_CODING_AGENT_DIR/skills/, but use directory-based skills for anything with bundled resources.

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

The model 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.

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. Share them with the user: [you don't have to use this exact language] "Here are a few test cases I'd like to try. Do these look right, or do you want to add more?" Then run them.

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.

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: Launch all runs (with-skill AND baseline) as one Pi orchestration batch

For each test case, run two Pi CLI executions — one with the skill, one baseline. Don't inspect or iterate on the with-skill output before the baseline exists. Either launch all configs from one small shell/Python orchestration command, or use tmux panes for longer runs. Sequential execution inside one script is fine for small evals; the important thing is that the iteration's run matrix is decided before you look at results.

Use pi --mode json --no-session for eval runs so the full event transcript is saved. Use consistent isolation flags across configurations. A clean default is --no-context-files --no-prompt-templates --no-skills; keep extensions enabled if they provide your model/provider, and only disable extensions when you know they are not needed.

With-skill run (force the skill for output-quality evaluation):

mkdir -p <workspace>/iteration-<N>/eval-<ID>/with_skill/run-1/outputs
/usr/bin/time -p -o <workspace>/iteration-<N>/eval-<ID>/with_skill/run-1/time.txt \
  pi --mode json --no-session --no-context-files --no-prompt-templates \
    --no-skills --skill <path-to-skill> \
    "/skill:<skill-name> Execute this task:
- Task: <eval prompt>
- Input files: <eval files if any, or none>
- Save outputs to: <workspace>/iteration-<N>/eval-<ID>/with_skill/run-1/outputs/
- Outputs to save: <what the user cares about — e.g., the .docx file, the final CSV>" \
  > <workspace>/iteration-<N>/eval-<ID>/with_skill/run-1/transcript.jsonl \
  2> <workspace>/iteration-<N>/eval-<ID>/with_skill/run-1/stderr.log

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

• Creating a new skill: no skill at all. Use --no-skills, omit --skill, and save to without_skill/run-1/outputs/.
• Improving an existing skill: the old version. Before editing, snapshot the skill (cp -r <skill-path> <workspace>/skill-snapshot/), then load that snapshot with --skill <workspace>/skill-snapshot and /skill:<skill-name>. Save to old_skill/run-1/outputs/.

Also save a human-readable transcript.md when practical (or keep the JSONL path and tell graders to read JSONL). Copy only the files the user should review into outputs/ so the viewer stays focused.

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 Pi process completes, save timing data to timing.json in the run directory. If you used /usr/bin/time -p, parse real seconds; if your Pi JSON events include usage/tokens in your environment, copy them too. Tokens are useful but optional — don't block the eval if only wall-clock time is available.

{
  "total_duration_seconds": 23.3,
  "duration_ms": 23332,
  "total_tokens": 0
}

Process each completion 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 — grade inline or launch a separate Pi CLI grading run 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. Generate the viewer as a standalone HTML file and open it:

   python <skill-creator-path>/eval-viewer/generate_review.py \
     <workspace>/iteration-N \
     --skill-name "my-skill" \
     --benchmark <workspace>/iteration-N/benchmark.json \
     --static <workspace>/iteration-N/review.html
   

   Then open it: open <workspace>/iteration-N/review.html

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

   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.

GENERATE THE EVAL VIEWER BEFORE evaluating inputs yourself. You want to get them in front of the human ASAP!

5. Tell the user something like: "I've opened the results in your browser. There are two tabs — 'Outputs' lets you click through each test case and leave feedback, 'Benchmark' shows the quantitative comparison. When you're done, come back here and let me know."

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.

---

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 agents all independently wrote similar helper scripts or took the same multi-step approach to something. If all 3 test cases resulted in the agent 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 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.

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

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 reviewer (inline or via a separate Pi CLI run) without telling it which is which, and let it judge quality. Then analyze why the winner won.

This is optional 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 the model 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 Pi 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 queries (8-10), think about coverage. You want different phrasings of the same intent — some formal, some casual. Include cases where the user doesn't explicitly name the skill or file type but clearly needs it. Throw in some uncommon use cases and cases where this skill competes with another but should win.

For the should-not-trigger queries (8-10), the most valuable ones are the near-misses — queries that share keywords or concepts with the skill but actually need something different. Think adjacent domains, ambiguous phrasing where a naive keyword match would trigger but shouldn't, and cases where the query touches on something the skill does but in a context where another tool is more appropriate.

The key thing to avoid: don't make should-not-trigger queries obviously irrelevant. "Write a fibonacci function" as a negative test for a PDF skill is too easy — it doesn't test anything. The negative cases should be genuinely tricky.

Step 2: Review with user

Present the eval set to the user for review using the HTML template:

1. Read the template from assets/eval_review.html
2. Replace the placeholders:
   • __EVAL_DATA_PLACEHOLDER__ -> the JSON array of eval items (no quotes around it — it's a JS variable assignment)
   • __SKILL_NAME_PLACEHOLDER__ -> the skill's name
   • __SKILL_DESCRIPTION_PLACEHOLDER__ -> the skill's current description
3. Write to a temp file (e.g., /tmp/eval_review_<skill-name>.html) and open it: open /tmp/eval_review_<skill-name>.html
4. The user can edit queries, toggle should-trigger, add/remove entries, then click "Export Eval Set"
5. The file downloads to ~/Downloads/eval_set.json — check the Downloads folder for the most recent version in case there are multiple (e.g., eval_set (1).json)

This step matters — bad eval queries lead to bad descriptions.

Step 3: Iterate on the description

Use the automated Pi scripts when the user wants quantitative triggering data and the Pi CLI is authenticated:

python <skill-creator-path>/scripts/run_eval.py \
  --eval-set /path/to/eval_set.json \
  --skill-path /path/to/skill \
  --model <pi-model-pattern>

python <skill-creator-path>/scripts/run_loop.py \
  --eval-set /path/to/eval_set.json \
  --skill-path /path/to/skill \
  --model <pi-model-pattern> \
  --results-dir /path/to/description-runs

run_eval.py creates temporary Pi skills with the candidate description, runs pi --mode json --no-session --no-skills --skill <temp-skill>, and counts the skill as triggered when Pi reads that temporary SKILL.md. run_loop.py uses pi -p to propose improved descriptions from the failures.

If the scripts are too slow, unavailable, or inappropriate for the user's setup, iterate manually:

1. Review which eval queries failed to trigger (or falsely triggered) with the current description
2. Rewrite the description to better cover the should-trigger cases while excluding the should-not-trigger cases
3. Test the updated description by asking the user to try the queries in a fresh Pi session or by running a small pi --mode json smoke test
4. Repeat until the triggering accuracy is satisfactory

How skill triggering works

Understanding the triggering mechanism helps design better eval queries. Pi includes available skills in the system prompt in XML format with each skill's name + description, and the model decides whether to read a skill based on that metadata. Users can force a skill with /skill:<name>, but automatic triggering still depends on the description. The important thing to know is that the model may skip skills for tasks it can easily handle on its own — simple, one-step queries like "read this PDF" may not trigger a skill even if the description matches perfectly, because the model can handle them directly with basic tools. Complex, multi-step, or specialized queries reliably trigger skills when the description matches.

This means your eval queries should be substantive enough that the model would actually benefit from consulting a skill. Simple queries like "read file X" are poor test cases — they won't trigger skills regardless of description quality.

Step 4: Apply the result

Update the skill's SKILL.md frontmatter with the improved description. Show the user before/after.

---

Updating an existing skill

The user might be asking you to update an existing skill, not create a new one. In this case:

• Preserve the original name. Note the skill's directory name and name frontmatter field -- use them unchanged.
• Edit the Pi skill location the user actually uses. Common locations are project .pi/skills/<skill-name>/SKILL.md and global $PI_CODING_AGENT_DIR/skills/<skill-name>/SKILL.md (in these dotfiles, ~/.config/pi/skills/<skill-name>/SKILL.md); Pi also supports .agents/skills/, ~/.agents/skills/, settings paths, packages, and explicit --skill paths. If managing skills via chezmoi, edit the corresponding source path (here, private_dot_config/pi/skills/).

---

Reference files

The agents/ directory contains instructions for specialized reviewer roles. Read them when you need to grade, compare, or analyze inline or via a separate Pi CLI run.

• agents/grader.md — How to evaluate assertions against outputs
• agents/comparator.md — How to do blind A/B comparison between two outputs
• agents/analyzer.md — How to analyze why one version beat another

The references/ directory has additional documentation:

• references/schemas.md — JSON structures for evals.json, grading.json, etc.

---

Repeating one more time the core loop here for emphasis:

• Figure out what the skill is about
• Draft or edit the skill
• Run test prompts with the skill loaded (via Pi CLI eval runs)
• With the user, evaluate the outputs:
  • Create benchmark.json and run eval-viewer/generate_review.py to help the user review them
  • Run quantitative evals
• Repeat until you and the user are satisfied
• Deliver the final skill to the user.

Pi has no built-in TodoList. If the work spans multiple turns, keep a short checklist in the workspace or in your response. Include: "Create evals JSON and run eval-viewer/generate_review.py so human can review test cases" so the human-review step is not skipped.

Good luck!