AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Skill Creator

skill-firstp1ck-pi-coding-agent-forge-skill-creator · by Firstp1ck

Use when drafting new reusable Pi/Agent Skills from repeated successful workflows, troubleshooting trajectories, notes, or PATCH.md files. Enforces reusability checks, writes disabled drafts, adds contract tests when possible, and asks before enablement.

No reviews yet
0 installs
15 views
0.0% view→install

Install

$ agentstack add skill-firstp1ck-pi-coding-agent-forge-skill-creator

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-firstp1ck-pi-coding-agent-forge-skill-creator)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
16d ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Skill Creator? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Skill Creator

Create Pi-native Agent Skills from successful repeated work without prematurely enabling unreviewed drafts.

Critical Correction to the Original Plan

Do not write drafts under ~/.pi/agent/skills/drafts/ by default. Pi recursively discovers SKILL.md files under ~/.pi/agent/skills/, so that path can accidentally expose drafts as available skills. Use this package's safe default instead:

~/.pi/agent/drafts/skills//SKILL.md

Only write under an auto-discovered skill root after explicit user approval and a review decision.

Lifecycle policy

Before drafting, enabling, publishing, or moving skills, read the bundled package policy at ../../../../references/SKILL-LIFECYCLE-POLICY.md when available. Use references/SKILL-PORTABILITY.md for portable skill authoring rules and references/SKILL-QUALITY-GUIDE.md for predictable, low-sprawl skill design. For the full upstream vocabulary and rationale, consult the vendored references/writing-great-skills/SKILL.md and references/writing-great-skills/GLOSSARY.md reference.

When to Use

Use this skill when the user wants to create a new skill from:

  • A repeated workflow that succeeded at least three times.
  • An expensive or high-risk workflow worth standardizing after one strong success.
  • Troubleshooting notes, LEARNINGS notes, or a PATCH.md that describe a reusable procedure.
  • A Pi package maintenance workflow that should become a portable Agent Skill.

Do not create a skill when the task is one-off, vague, unverified, or better captured as a short memory note.

Reusability Gate

Before drafting, confirm one of these is true:

  1. The workflow has been reused successfully at least three times.
  2. The workflow is expensive/risky enough that standardizing it prevents meaningful future cost.
  3. The workflow is strategically likely to recur across packages/projects.

If none is true, recommend a note or LEARNINGS entry instead of a skill.

Workflow

  1. Collect source evidence
  • Read the successful trajectory, notes, PATCH.md, or relevant files.
  • Verify the source describes an outcome that actually worked.
  1. Confirm reusability
  • Record the reusability class and concise evidence.
  • Stop if the workflow is not reusable enough.
  • For portable skills, follow references/SKILL-PORTABILITY.md.
  1. Design for predictability
  • Decide whether the draft should be model-invoked or user-invoked.
  • Identify leading words, trigger branches, should-trigger examples, and should-not-trigger examples.
  • Keep ordered actions inline; disclose long or branch-specific reference through references/*.md.
  • Ensure each workflow step has a checkable completion criterion.
  • Prune no-ops, duplicated meanings, stale source sediment, and sprawl using references/SKILL-QUALITY-GUIDE.md; consult references/writing-great-skills/ when the concise guide is insufficient.
  1. Draft the skill
  • Prefer the skill_create_draft tool when available.
  • Otherwise run node ./scripts/skill_create_draft.mjs from this skill directory.
  • Default to ~/.pi/agent/drafts/skills//.
  1. Add fixtures/tests when possible
  • Use --with-tests or tool parameter withTests: true for a basic contract test.
  • Include only sanitized, non-secret source snippets as fixtures.
  1. Validate
  • Run the generated contract test.
  • Run skill_eval_run /SKILL.md if the skill evaluator is installed.
  • If the evaluator is absent, use the built-in draft validation output and clearly report that the external evaluator is unavailable.
  1. Review before enabling
  • Do not symlink, install, or add the draft to settings automatically.
  • Ask the user before moving a draft into ~/.pi/agent/skills/ or enabling a package.

Tool Usage

Draft from notes

node ./scripts/skill_create_draft.mjs \
  --name example-repeatable-workflow \
  --source-notes /tmp/example-successful-trajectory.md \
  --reusability repeated-3-plus \
  --reusability-evidence "Used successfully for three similar tasks." \
  --invocation model \
  --leading-word repeatable \
  --branch "successful workflow reuse" \
  --with-tests

Draft package skeleton

node ./scripts/skill_create_draft.mjs \
  --name example-repeatable-workflow \
  --source-notes /tmp/example-successful-trajectory.md \
  --reusability strategic-reuse \
  --reusability-evidence "Expected to recur across Pi package maintenance work." \
  --package-skeleton \
  --output /tmp/pi-skill-example-repeatable-workflow \
  --with-tests

Draft Quality Contract

Every generated SKILL.md must include:

  • Valid frontmatter with lowercase hyphenated name and a specific description.
  • Deliberate invocation mode, including disable-model-invocation: true for user-invoked drafts.
  • ## When to Use with concrete trigger and non-trigger guidance.
  • ## Invocation Design with leading words, branches, and trigger examples.
  • ## Workflow with checkable completion criteria for each step.
  • ## References for progressive disclosure decisions.
  • ## Verification.
  • ## Quality Review covering no-ops, duplication, sediment, and sprawl.
  • ## Safety and Failure Modes.
  • No hardcoded private home paths unless the draft is explicitly marked Pi-local.
  • A clear note that the draft is not enabled automatically.

Verification

For this package:

cd 
npm test

For a generated draft with tests:

python3 -m unittest discover -s /tests -p 'test_*.py'

If the skill evaluator is installed:

skill_eval_run /SKILL.md

Safety and Failure Modes

  • Never auto-enable a generated draft.
  • Never write drafts into Pi's discovered skill roots without explicit approval.
  • Do not store secrets in fixtures, generated skills, or source excerpts.
  • Sanitize private paths such as /home//... to ~/... for portable drafts.
  • Stop and ask for clarification when the source artifact does not contain a clear, successful workflow.
  • Prefer a memory note or LEARNINGS note when a workflow is not reusable enough.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.