skill-builder

A guide for creating a new Skill, meaning a reusable set of instructions that teaches an agent how to handle a type of task. It keeps the new Skill consistent with the framework's existing structure and conventions.

In plain words
What is it for?
Use it to turn a described need or an approved evolution proposal into a new modular Skill that can be added to the framework.
Why use it?
It removes the uncertainty of deciding how a new Skill should be organized and written. It also helps avoid using this workflow for editing an existing Skill or recording feedback.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/zxpmail/reqforge/skill-builder
Any agent
npx skills add zxpmail/ReqForge --skill skill-builder
Clone the repo
git clone --depth 1 https://github.com/zxpmail/ReqForge

Made for: Claude Code, Codex.

Per session 0 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,441 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.02441
Opus 5 $0.00000 $0.01221
Sonnet 5 $0.00000 $0.00488
Haiku 4.5 $0.00000 $0.00244

Measured 2d ago against content hash 26b397b02109, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

skill-builder scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

adapters/claude-code/.claude/skills/skill-builder/SKILL.md · 193 lines

How it starts

The opening of the file, as written. The whole thing — 193 lines — stays where its author put it; the contents beside it link to each section on GitHub.


name: skill-builder description: Used when the user wants to create a new Skill, or when an EVOLUTION.md proposal auto-generates a new Skill. Creates a structurally consistent new Skill following the framework's modular conventions. version: 1.0.0 updated: 2026-05-26 requires: []

[Task] Create a new Skill that conforms to framework conventions based on the user's described needs or an EVOLUTION.md fourth-layer proposal. Ensure the new Skill shares the same structure, unified style, and plug-and-play modularity as existing Skills.

[Not For] - Modifying existing Skills -> edit the SKILL.md directly instead - Evolving rules from feedback patterns -> use /evolution-engine instead - Recording feedback about a Skill -> use /feedback-writer instead

[Dependency Check] Required: None (this Skill does not depend on external files)

Optional:
- Related records in ../../feedback/ -> if from an EVOLUTION.md proposal, read the original feedback to understand the need's background

[First Principles] → references/first-principles.md 核心:Template First / Reference Existing / Minimum Necessary / Web-First。

[Output Style] → references/output-style.md Tone: Architect explaining a blueprint — structured, precise, prescriptive.

[File Structure] skill-builder/ ├── SKILL.md # Main Skill definition (this file) ├── templates/ │ └── skill-template.md # Skeleton template for new Skills └── references/ ├── first-principles.md ├── output-style.md ├── workflow.md ├── dimension-checklist.md └── anti-rationalization.md

[Dimension Checklist] See references/dimension-checklist.md for the full dimension checklist.

Must-have dimensions:
- **Decidable Triggers**: description specifies when to use AND when not
- **Workflow Executability**: each step specifies concrete action
- **Gotchas from Practice**: ≥3 specific failure points with guidance
- **Boundary Clarity**: [Not For] section with explicit exclusions
- **Dependency Check**: required deps have failure guidance

[Anti-Rationalization Checklist] → references/anti-rationalization.md 遇 skipping template / skipping reference / TBD placeholders 时读取。

[Gotchas] Skipping template: "I know the structure well enough" — read the template anyway. Every time you skip, you'll miss something: a section heading, a required field, or the consistent format. Not cross-referencing existing Skills: Writing in a different style from the rest of the codebase. Always read 1-2 existing Skills before creating a new one. Consistency matters for maintainability. Empty sections: "To be filled later" is technical debt. If a section isn't needed, don't include it. If it IS needed, fill it now. Empty sections in a Skill cause confusion when the Skill is invoked. Missing Gotchas: You're building a Skill that WILL accumulate failure points. If you don't leave a [Gotchas] section, where will those lessons go? Nowhere — they'll be repeated. Skipping Skill eval: Shipping without triggers.json / cases.json — you won't know if description misfires or outputs regress. Run pnpm skill-eval after every meaningful SKILL.md change.

[Quality Rubric] 16-item, 32-point scoring system. Ship threshold: ≥ 24 with no critical item scoring 0.

Read the full file on GitHub · 193 lines

Files

What ships with it

8 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 193 lines · 0 tokens per session scan A 26b397b02109

Subscribe to this mod's changes

skill-builder is a skill published in the GitHub repository zxpmail/ReqForge (18 stars, last pushed 12d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,441 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

openrig-operating-model

Use when you do not know WHERE something belongs: you learned a lesson and are unsure which file takes it; you are about to create a new doc, folder, or convention; you are asking "should this be a skill, a note, or a chain file?"; you inherited a seat and want to know what SHOULD exist; or you are about to duplicate…

mvschwarz/openrig · 113 tokens

memex-retro

Save insights from completed tasks to Zettelkasten memory.

iamtouchskyer/memex · 17 tokens

direction-adjust

Capture a small-to-medium course correction the user realized mid-task (not from external review), record it as a numbered D-N entry in DECISIONS.md, update TASKSTATE.md to reflect the adjusted direction, and route back to the appropriate command. Use when you are mid-task (any phase past discovery) and realize the…

Mozurok/fhorja.dev · 201 tokens

harvest-session-learnings

Scan the current working session and the active task's artifacts for reusable, generalizable lessons (what was tried, what failed and why, what surprised us, what the next task should do differently) and propose anchored entries to append to the task's LEARNINGS.md, the produce-side counterpart to the ADR-0017 consume…

Mozurok/fhorja.dev · 198 tokens

doc-governance

Audit and converge a repository's human documentation using the governance model proven in practice: single source of truth, lifecycle-tiered docs (current truth / lessons / validation), project content kept out of pipeline docs, dead docs deleted instead of archived, and honest evidence labels. Use when docs feel…

manhua-man/harness-coding-protocol · 85 tokens

javascript-mastery

33+ essential JavaScript concepts every developer should know, inspired by 33-js-concepts.

KunanonJ/ai-skills-hub · 38 tokens