--- name: skill-creator description: | Use this skill when the user wants to create or improve a WrongStack skill (SKILL.md). Triggers: user says "create a skill", "new skill", "add a skill", "skill definition". version: 1.3.0 required-capabilities: [filesystem.write, runtime.admin] required-tools: [bash, skill] --- # Skill Creator — WrongStack ## Overview Guides the creation of new WrongStack skills. A skill is a Markdown file with YAML frontmatter — the full description tells the model when to use it. You are the wizard: ask questions, validate answers, write the file. Use the `/skill-gen` sub-commands to do the mechanical parts (validation, scaffolding) deterministically. ## Rules 1. Write a complete `description` (1–1024 characters) explaining what the skill does and when to use it. The full description is disclosed to the model. 2. Name must be kebab-case: `my-skill`, `docker-deploy` — lowercase, hyphens only. 3. Skills live in `.wrongstack/skills//SKILL.md` (project level). 4. After the trigger sentence, add `Triggers: user says "X", "Y", "Z".`. 5. Content must be actionable — rules, patterns, anti-patterns, not just prose. 6. End with "Skills in scope" listing related skills for delegation. 7. Don't let skill names collide with existing skills. ## Authoring commands (use these) The `/skill-gen` command is a toolkit — pick the right sub-command for the job instead of writing files by hand: | Sub-command | When to use | |---|---| | `/skill-gen` (bare) | Open-ended creation: you (the agent) ask questions one at a time, then write the file. Best for nuanced skills. | | `/skill-gen skeleton --desc "..." --trigger a,b` | Quick scaffold: generates a valid SKILL.md skeleton the user edits. Use when the name + trigger are already known. | | `/skill-gen from-prompt ""` | Turn an existing prompt/instruction into a skill draft. Use when the user hands you a prompt and says "make this a skill". | | `/skill-gen validate ` | Validate an existing skill file, or check a proposed name before writing. | | `/skill-gen view ` | Read-only: show a skill's body. | | `/skill-gen edit ` | Open the skill in `$EDITOR`/`$VISUAL`. | | `/skill-gen list` | List skills with their source layer. | After writing a skill with the wizard flow, run `/skill reload`, then `/skill-gen validate ` to check its YAML, name, field limits, and body. Use `/skill use ` to explicitly apply it. Skeletons and prompt drafts refresh the loader automatically. ## Patterns ### Do ```markdown --- name: docker-deploy description: | Use this skill when deploying Docker containers to a production cluster. Triggers: user says "docker", "container", "deploy", "dockerfile", "image". version: 1.0.0 --- # Docker Deploy — WrongStack ## Overview ... ## Rules ... ## Patterns ... ## Skills in scope ``` ### Don't ```markdown --- name: MySkill # ❌ PascalCase name: my_skill # ❌ underscore description: | This skill is about Docker. # ❌ no trigger sentence --- ``` ## Skill format Every skill is a Markdown file with YAML frontmatter: ```markdown --- name: my-skill-name description: | Use this skill when . Triggers: user says "keyword", "another keyword". version: 1.0.0 --- # Skill Title ## Overview What this skill does. ## Rules - Rule 1 - Rule 2 ## Patterns ### Do \`\`\`ts // good example \`\`\` ### Don't \`\`\`ts // bad example \`\`\` ## Workflow 1. Step one 2. Step two ``` ## File structure A skill is a directory containing `SKILL.md` plus optional resource subdirectories (the agentskills.io layout): ``` / SKILL.md ← required: metadata + instructions scripts/ ← optional: executable code (run via bash) references/ ← optional: docs loaded on demand (REFERENCE.md, …) assets/ ← optional: templates, data, snippets … ← any other subdirectories ``` Skills live under these paths (priority order, first-seen wins by name): 1. **Project**: `/.wrongstack/skills//` 2. **Project foreign**: `/.claude/skills//`, `/.{codex,cursor,agents,…}/skills//` 3. **User profile**: `~/.wrongstack/profiles//skills//` 4. **User foreign**: `~/.claude/skills//`, `~/.{codex,cursor,agents,…}/skills//` 5. **Bundled**: `packages/core/skills//` (read-only, core team) For user-created skills: always use path 1 (project level). ## Resource files (scripts / references / assets) Bundled resources are NOT injected into the prompt — the agent loads them on demand via the `skill` tool (agentskills.io progressive disclosure, tier 3): - `skill({ name: "" })` → lists every bundled file (scripts/, references/, assets/, any subdir, recursively). - `skill({ name: "", resource: "references/REF.md" })` → returns that file's content. Scripts come back with an absolute path so the agent runs them via `bash`. Keep `SKILL.md` under ~500 lines; move deep reference material into `references/`. Reference files with relative paths from the skill root. Scripts must be self-contained and safe to run. Only add the subdirectories a skill actually needs — empty directories don't persist in git, so create a file (e.g. `scripts/README.md`) if you want the directory tracked. ## Workflow 1. **Ask the name** — suggest kebab-case, validate format 2. **Ask the trigger** — "What situation should activate this skill?" 3. **Ask the coverage** — what rules, patterns, workflows? 4. **Generate the SKILL.md** — write to `.wrongstack/skills//SKILL.md` 5. **Confirm** — show the path, remind them to use `/skill` to list skills ## Validation Checklist Before writing the file, verify: - [ ] Name is valid kebab-case - [ ] Name doesn't collide with existing skills - [ ] Description has a clear trigger sentence - [ ] Content is actionable (rules, patterns, not just prose) - [ ] File will be placed in `.wrongstack/skills/` ## Out of scope - **Don't write skills without running `/skill-gen validate ` first.** A name that collides with an existing skill or breaks the kebab-case rule is a wall hit after the file is written. Validate first, always. - **Don't ship a skill with a vague description.** "This skill is about Docker" matches nothing. First sentence = trigger; rest = `Triggers: user says "..."` listing concrete keywords. - **Don't name skills in PascalCase or with underscores.** `MySkill` and `my_skill` are wrong; only `my-skill` (kebab-case) loads. The loader rejects the others. - **Don't place project skills outside `.wrongstack/skills//`.** Bundled, user-profile, and foreign paths are for other owners. User-created skills always go in the project directory. - **Don't write prose-only skills.** Rules, patterns, anti-patterns, code examples — the content has to be actionable. Prose without a checkable rule does not constrain the model. - **Don't skip "Skills in scope".** Without the hand-off list, the model doesn't know where to delegate adjacent questions. The list is what makes the skill part of a system. - **Don't forget to bump the version on structural change.** New section, scope change, rule change → major or minor bump. Patch only for wording. - **Don't move work into the skill that belongs in `/skill-gen`.** The sub-commands are the deterministic layer: validation, scaffolding, from-prompt conversion. The wizard is for the parts they can't do. ## Before returning - [ ] Name is kebab-case; `/skill-gen validate ` passed - [ ] Name doesn't collide with bundled, project, or user-profile skills - [ ] First sentence of `description` is a concrete trigger; trigger keywords follow - [ ] File path is `.wrongstack/skills//SKILL.md` for project skills - [ ] Content is actionable: rules, patterns, anti-patterns, code examples - [ ] "Out of scope" lists what the skill is NOT for and where to hand off - [ ] "Before returning" gives the model a mechanical completion check - [ ] "Skills in scope" names the hand-off targets and the reason for each - [ ] Version bumped appropriately; change recorded in CHANGELOG.md - [ ] `/skill-gen validate ` re-run after writing; loads cleanly ## Skills in scope - `prompt-engineering` — for crafting the skill description and prompt text - `git-flow` — for committing the new skill file - `output-standards` — for standardized `` formatting