[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M71HG5CSP04F29HRAG7XVYKG":3,"skill-same:01M71HG5CSP04F29HRAG7XVYKG":113},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":25,"publishedAt":46,"linkOnly":47,"markdownContent":48,"markdownBody":49,"frontMatter":50,"refName":54,"commitSha":55,"skillDirectory":56,"forksCount":57,"repositoryHtmlUrl":58,"repositoryHomepageUrl":6,"ownerHtmlUrl":59,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":60,"install":61,"resources":74},"01M71HG5CSP04F29HRAG7XVYKG","skill-forge",null,"Designs, authors, validates, and optimizes new Agent Skills from scratch. Use when creating a new skill, packaging a workflow as a skill, improving a skill's triggering accuracy, or evaluating whether a skill improves output quality. Covers the full lifecycle: DISCOVER, DESIGN, AUTHOR, VALIDATE, OPTIMIZE, DELIVER. Produces skills that follow the open SKILL.md format (YAML frontmatter + Markdown body, progressive disclosure, validation loops). EN triggers: \u002Fforge, create a skill, write a skill, build a skill, new skill, author a skill, package this as a skill, validate a skill, improve skill triggering, optimize skill description, evaluate skill quality. PT triggers: \u002Fforge, criar uma skill, escrever uma skill, construir uma skill, nova skill, empacotar como skill, validar skill, melhorar disparo, otimizar descrição de skill. Do NOT use for: discovering which existing skill to apply (use a dispatch\u002Fmeta skill), planning feature implementation, or designing software architecture.","runecraftai\u002Farcanum","runecraftai","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F279098826?v=4","01M71HBGS1RCAJ490RAYST7M00","01M71HBGT1CB94MNS6T6SGAZKW",1,"CC-BY-4.0","https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fskill-forge\u002FSKILL.md",[17,22],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000015","视觉设计","SCENARIO","使用场景",{"categoryNo":23,"categoryName":24,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000023","Skill与智能体开发",[26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45],"skill forge","forge","full","new skill","a skill","software","write","de","create","architecture","design","open","improve","Meta","packaging","pt","feature implementation","new agent","which","deliver","2026-11-10 10:12:30",false,"---\nname: skill-forge\ndescription: >\n  Designs, authors, validates, and optimizes new Agent Skills from scratch. Use when creating\n  a new skill, packaging a workflow as a skill, improving a skill's triggering accuracy, or\n  evaluating whether a skill improves output quality. Covers the full lifecycle: DISCOVER,\n  DESIGN, AUTHOR, VALIDATE, OPTIMIZE, DELIVER. Produces skills that follow the open\n  SKILL.md format (YAML frontmatter + Markdown body, progressive disclosure, validation loops).\n  EN triggers: \u002Fforge, create a skill, write a skill, build a skill, new skill, author a skill,\n  package this as a skill, validate a skill, improve skill triggering, optimize skill description,\n  evaluate skill quality.\n  PT triggers: \u002Fforge, criar uma skill, escrever uma skill, construir uma skill, nova skill,\n  empacotar como skill, validar skill, melhorar disparo, otimizar descrição de skill.\n  Do NOT use for: discovering which existing skill to apply (use a dispatch\u002Fmeta skill),\n  planning feature implementation, or designing software architecture.\nlicense: CC-BY-4.0\nmetadata:\n  version: 1.0.0\n---\n\n# skill-forge\n\n> Meta-skill for creating new Agent Skills from scratch. Produces publishable, validated, trigger-tested skills that compose with the rest of the catalog.\n\n```\nDISCOVER → DESIGN → AUTHOR → VALIDATE → OPTIMIZE → DELIVER\n```\n\nEach phase has explicit exit criteria. Move sequentially. Do not skip to AUTHOR before completing DISCOVER and DESIGN — bad frontmatter is the most common reason skills misfire.\n\n## Critical rules (read before acting)\n\n- **The open SKILL.md format is the source of truth.** This skill encodes the format rules (frontmatter schema, naming, progressive disclosure, folder layout). When in doubt, follow what this skill's own validator enforces — it is the single reference implementation.\n- **The description field controls triggering.** It is the only thing the agent sees at startup. Write it last but treat it as the most important field. If triggering is wrong, nothing else matters.\n- **Skills are for agents, not humans.** The `SKILL.md` body is agent instructions. Human-facing docs (README, changelogs) belong outside the skill folder.\n- **Never fabricate domain knowledge.** Build skills from real expertise — your own run history, internal docs, runbooks, schemas, code review comments. Generic \"best practices\" skills have no edge.\n\n---\n\n## Phase 1: DISCOVER\n\n**Goal:** Build a precise mental model of the skill before writing a single line.\n\nAsk one cluster at a time (don't dump). Capture answers in your working memory; you'll reuse them in DESIGN.\n\n### 1.1 What workflow are we packaging?\n\n- What does the user do today, step by step?\n- Where does it go wrong without the skill (inconsistency, forgotten steps, wasted re-explaining)?\n- What is the \"good output\" they expect (specific, observable)?\n\n### 1.2 Use cases\n\nNail down 2-3 concrete ones:\n\n```\nUse Case: [Name]\nTrigger phrase: \"user would say…\"\nSteps: 1. … 2. … 3. …\nTools: built-in or MCP\nSuccess: [specific output or state]\n```\n\nIf vague, offer examples to react to. Concrete > abstract.\n\n### 1.3 Category\n\n| Category | When to choose |\n|---|---|\n| Document & Asset Creation | Consistent output generation (reports, slides, code, diagrams) |\n| Workflow Automation | Multi-step processes with methodology (planning, onboarding, deploys) |\n| MCP Enhancement | Workflow guidance on top of MCP tool access |\n| Domain Intelligence | Specialized knowledge (compliance, finance rules, security checks) |\n\n### 1.4 Out of scope (explicit)\n\nList adjacent things this skill must NOT do. Add them to the `Do NOT use for…` clause of the description.\n\n**Exit:** 2-3 use cases, category, out-of-scope list.\n\n---\n\n## Phase 2: DESIGN\n\n**Goal:** Structural decisions before any writing.\n\n### 2.1 Pick a pattern\n\n| Pattern | When |\n|---|---|\n| Sequential Workflow | Steps have dependencies, ordering is load-bearing |\n| Multi-MCP Coordination | Spans multiple external services |\n| Iterative Refinement | Quality improves through review-fix cycles |\n| Context-Aware Selection | Best tool depends on input properties |\n| Domain-Specific Intelligence | Expert rules drive correct execution |\n\nMost skills combine. Identify the **primary** pattern.\n\n### 2.2 Folder layout\n\n```\nskill-name\u002F\n├── SKILL.md           # Required. Target \u003C500 lines, \u003C5000 tokens.\n├── scripts\u002F           # Only if deterministic checks needed\n├── references\u002F        # Only if deep content >100 lines\n├── assets\u002F            # Only if templates\u002Fstatic resources are used in output\n└── evals\u002F             # Optional, for output-quality evaluation\n    └── evals.json\n```\n\nDecision rules:\n- Logic that must be deterministic → `scripts\u002F`\n- Reference content >100 lines → `references\u002F`\n- Templates\u002Fimages reused in output → `assets\u002F`\n- Everything else → keep in SKILL.md\n\n### 2.3 Map progressive disclosure\n\n| Level | What | Token budget |\n|---|---|---|\n| L1 Frontmatter | `name` + `description` | ~100 words |\n| L2 SKILL.md body | Core workflow, steps, examples | \u003C500 lines |\n| L3 Linked files | Deep ref, API docs, large examples | As needed |\n\nSKILL.md must tell the agent **when** to read each linked file. \"Read `references\u002Fspec.md` if X\" beats \"see references\u002F\".\n\n### 2.4 Draft the description (placeholder, finalize in OPTIMIZE)\n\nFormat: `[What] + [Use when, with trigger phrases] + [Do NOT use for]`. ≤1024 chars total.\n\n**Exit:** Pattern, folder layout, L1\u002FL2\u002FL3 map, description draft.\n\n---\n\n## Phase 3: AUTHOR\n\n**Goal:** Write the skill with precision.\n\n### 3.1 Frontmatter hard rules\n\n```yaml\n---\nname: kebab-case-name           # Must match folder. 1-64 chars, [a-z0-9-], no leading\u002Ftrailing\u002Fconsecutive hyphens.\ndescription: >                   # Folded multiline OK; ≤1024 chars; no XML angle brackets.\n  What it does. Use when [triggers]. Do NOT use for [exclusions].\nlicense: CC-BY-4.0\nmetadata:\n  version: 1.0.0\n  author: your-name-or-org\n---\n```\n\nForbidden:\n- Spaces or capitals in `name` or folder\n- `claude` \u002F `anthropic` reserved terms\n- `SKILL.MD`, `Skill.md`, `skill.md` (must be exactly `SKILL.md`)\n- XML angle brackets `\u003C` `>` in description\n- Description > 1024 characters\n\n### 3.2 Body writing principles\n\n- Imperative form (\"Run X\", \"Check Y\"). Specific > verbose.\n- Critical rules at the top, not buried in the middle.\n- 2-3 concrete examples (input → actions → result).\n- Reference files with explicit load conditions.\n- For deterministic checks, prefer scripts over prose.\n- Do not wrap prose at arbitrary column widths; let sentences flow.\n- Code blocks may wrap for readability.\n\n### 3.3 Supporting files\n\nFor each `references\u002F*.md` or `scripts\u002F*`:\n- Reference it from SKILL.md with a clear WHEN clause.\n- For files >300 lines, add a Table of Contents.\n- Scripts: see [scripts-guide](references\u002Fscripts-guide.md) for PEP 723, agentic design, --help, exit codes.\n\n**Exit:** SKILL.md drafted, all hard rules satisfied, supporting files written.\n\n---\n\n## Phase 4: VALIDATE\n\n**Goal:** Catch structural and spec errors before delivery.\n\n### 4.1 Run the bundled validator\n\n```bash\npython3 scripts\u002Fvalidate.py \u003Cskill-folder>\n# or, with PEP 723 inline dependencies via uv:\nuv run --no-project scripts\u002Fvalidate.py \u003Cskill-folder>\n```\n\nThe validator checks: folder kebab-case, SKILL.md exists, frontmatter YAML valid, `name` rules, `description` rules, body line count, examples present, references linked, no XML in description.\n\nPass = exit code 0. Warnings are allowed. Errors are not.\n\n### 4.2 Spec re-check\n\nConfirm against the full format spec (see [references\u002Fspec.md](references\u002Fspec.md)):\n- `compatibility` field if the skill needs specific environment\n- `allowed-tools` field if you want to pre-approve specific tools\n- File references are relative to the skill root, one level deep preferred\n\n### 4.3 Trigger sanity\n\nMentally run 5-10 short prompts. Should trigger; should NOT trigger (near-misses). If the description is too narrow, broaden it; if too broad, add exclusions. Save these prompts — they seed the eval in OPTIMIZE.\n\nFor the full eval-driven optimization, proceed to OPTIMIZE.\n\n**Exit:** Validator PASS + spec check + trigger sanity.\n\n---\n\n## Phase 5: OPTIMIZE\n\n**Goal:** Make the skill trigger reliably and prove it improves output quality.\n\nTwo eval loops, run in this order:\n\n### 5.1 Description triggering eval\n\nOpen [description-optimization](references\u002Fdescription-optimization.md).\n\nWrite ~20 eval queries (8-10 should-trigger, 8-10 should-NOT-trigger), emphasizing near-misses. Split 60\u002F40 into `train_queries.json` and `validation_queries.json`. Run the trigger script 3x per query. Iterate on the description (using only the train set) until stable. Select the best iteration by validation pass rate.\n\n### 5.2 Output-quality eval (optional, recommended for workflow skills)\n\nOpen [output-evaluation](references\u002Foutput-evaluation.md).\n\nBuild `evals\u002Fevals.json` with 2-3 test cases. Run each twice (with-skill, without-skill) into `iteration-1\u002F`. Add assertions, grade, aggregate to `benchmark.json`. Iterate the skill body based on failed assertions and human review. Stop when pass rate plateaus or feedback is empty.\n\n**Exit:** Description eval pass rate ≥0.5 on validation set; output eval (if run) shows skill adds value.\n\n---\n\n## Phase 6: DELIVER\n\n**Goal:** Install + register.\n\n### 6.1 Local install\n\nDrop the skill into the right skills directory for the target agent:\n\n| Agent | Path |\n|---|---|\n| VS Code + Copilot | `.agents\u002Fskills\u002F` |\n| Claude Code | `.claude\u002Fskills\u002F` |\n| Cursor | `.cursor\u002Fskills\u002F` |\n| OpenCode | `.opencode\u002Fskill\u002F` |\n\nOptionally add a slash command to the agent's command file (e.g. `\u002Fforge`) so the trigger works without typing a long phrase.\n\n### 6.2 For catalog publication\n\n1. Add a row to the catalog's index (package README) with version, description, main trigger, and link to the skill's README.\n2. Create a release note in whatever format the catalog uses (changeset, changelog entry, release tag, etc.).\n3. Bump the skill's own `metadata.version` to match the release.\n4. Publish through whatever distribution mechanism the catalog uses (npm package, git tag, registry upload).\n\n### 6.3 For distribution elsewhere\n\nKeep the skill self-contained. Make sure the skill folder is at a path the target agent scans. Include a top-level human-facing README alongside the skill folder (not inside it).\n\n**Exit:** Skill installed, registered, and a test prompt verifies triggering in the target agent.\n\n---\n\n## Bundled resources\n\n- **[references\u002Fspec.md](references\u002Fspec.md)** — the open SKILL.md format reference.\n- **[references\u002Fauthoring-patterns.md](references\u002Fauthoring-patterns.md)** — best practices: gotchas, templates, checklists, validation loops, plan-validate-execute, calibrating control.\n- **[references\u002Fdescription-optimization.md](references\u002Fdescription-optimization.md)** — trigger eval methodology with train\u002Fval split.\n- **[references\u002Foutput-evaluation.md](references\u002Foutput-evaluation.md)** — output-quality eval: with-skill vs without-skill, assertions, benchmark.\n- **[references\u002Fscripts-guide.md](references\u002Fscripts-guide.md)** — using scripts in skills: PEP 723, agentic script design, --help, exit codes.\n- **[scripts\u002Fvalidate.py](scripts\u002Fvalidate.py)** — stdlib-only validator; `--json-out` for agentic consumption.\n- **[assets\u002FSKILL.template.md](assets\u002FSKILL.template.md)** — blank template for new skills.\n\n## Anti-patterns to avoid\n\n- ❌ Vague description: \"Helps with documents.\"\n- ❌ Verbose body the agent will skim.\n- ❌ No examples.\n- ❌ Wrong `SKILL.md` casing.\n- ❌ Spaces or capitals in folder or name.\n- ❌ `claude` or `anthropic` in the name.\n- ❌ XML angle brackets in description.\n- ❌ SKILL.md > 500 lines (move detail to references\u002F).\n- ❌ Description > 1024 chars.\n- ❌ Claiming to be the only skill loaded (skills compose).\n- ❌ Re-explaining what the agent already knows (PDFs, HTTP, etc.).\n- ❌ Referencing internal projects, packages, or repos by name in the skill body.\n","# skill-forge\n\n> Meta-skill for creating new Agent Skills from scratch. Produces publishable, validated, trigger-tested skills that compose with the rest of the catalog.\n\n```\nDISCOVER → DESIGN → AUTHOR → VALIDATE → OPTIMIZE → DELIVER\n```\n\nEach phase has explicit exit criteria. Move sequentially. Do not skip to AUTHOR before completing DISCOVER and DESIGN — bad frontmatter is the most common reason skills misfire.\n\n## Critical rules (read before acting)\n\n- **The open SKILL.md format is the source of truth.** This skill encodes the format rules (frontmatter schema, naming, progressive disclosure, folder layout). When in doubt, follow what this skill's own validator enforces — it is the single reference implementation.\n- **The description field controls triggering.** It is the only thing the agent sees at startup. Write it last but treat it as the most important field. If triggering is wrong, nothing else matters.\n- **Skills are for agents, not humans.** The `SKILL.md` body is agent instructions. Human-facing docs (README, changelogs) belong outside the skill folder.\n- **Never fabricate domain knowledge.** Build skills from real expertise — your own run history, internal docs, runbooks, schemas, code review comments. Generic \"best practices\" skills have no edge.\n\n---\n\n## Phase 1: DISCOVER\n\n**Goal:** Build a precise mental model of the skill before writing a single line.\n\nAsk one cluster at a time (don't dump). Capture answers in your working memory; you'll reuse them in DESIGN.\n\n### 1.1 What workflow are we packaging?\n\n- What does the user do today, step by step?\n- Where does it go wrong without the skill (inconsistency, forgotten steps, wasted re-explaining)?\n- What is the \"good output\" they expect (specific, observable)?\n\n### 1.2 Use cases\n\nNail down 2-3 concrete ones:\n\n```\nUse Case: [Name]\nTrigger phrase: \"user would say…\"\nSteps: 1. … 2. … 3. …\nTools: built-in or MCP\nSuccess: [specific output or state]\n```\n\nIf vague, offer examples to react to. Concrete > abstract.\n\n### 1.3 Category\n\n| Category | When to choose |\n|---|---|\n| Document & Asset Creation | Consistent output generation (reports, slides, code, diagrams) |\n| Workflow Automation | Multi-step processes with methodology (planning, onboarding, deploys) |\n| MCP Enhancement | Workflow guidance on top of MCP tool access |\n| Domain Intelligence | Specialized knowledge (compliance, finance rules, security checks) |\n\n### 1.4 Out of scope (explicit)\n\nList adjacent things this skill must NOT do. Add them to the `Do NOT use for…` clause of the description.\n\n**Exit:** 2-3 use cases, category, out-of-scope list.\n\n---\n\n## Phase 2: DESIGN\n\n**Goal:** Structural decisions before any writing.\n\n### 2.1 Pick a pattern\n\n| Pattern | When |\n|---|---|\n| Sequential Workflow | Steps have dependencies, ordering is load-bearing |\n| Multi-MCP Coordination | Spans multiple external services |\n| Iterative Refinement | Quality improves through review-fix cycles |\n| Context-Aware Selection | Best tool depends on input properties |\n| Domain-Specific Intelligence | Expert rules drive correct execution |\n\nMost skills combine. Identify the **primary** pattern.\n\n### 2.2 Folder layout\n\n```\nskill-name\u002F\n├── SKILL.md           # Required. Target \u003C500 lines, \u003C5000 tokens.\n├── scripts\u002F           # Only if deterministic checks needed\n├── references\u002F        # Only if deep content >100 lines\n├── assets\u002F            # Only if templates\u002Fstatic resources are used in output\n└── evals\u002F             # Optional, for output-quality evaluation\n    └── evals.json\n```\n\nDecision rules:\n- Logic that must be deterministic → `scripts\u002F`\n- Reference content >100 lines → `references\u002F`\n- Templates\u002Fimages reused in output → `assets\u002F`\n- Everything else → keep in SKILL.md\n\n### 2.3 Map progressive disclosure\n\n| Level | What | Token budget |\n|---|---|---|\n| L1 Frontmatter | `name` + `description` | ~100 words |\n| L2 SKILL.md body | Core workflow, steps, examples | \u003C500 lines |\n| L3 Linked files | Deep ref, API docs, large examples | As needed |\n\nSKILL.md must tell the agent **when** to read each linked file. \"Read `references\u002Fspec.md` if X\" beats \"see references\u002F\".\n\n### 2.4 Draft the description (placeholder, finalize in OPTIMIZE)\n\nFormat: `[What] + [Use when, with trigger phrases] + [Do NOT use for]`. ≤1024 chars total.\n\n**Exit:** Pattern, folder layout, L1\u002FL2\u002FL3 map, description draft.\n\n---\n\n## Phase 3: AUTHOR\n\n**Goal:** Write the skill with precision.\n\n### 3.1 Frontmatter hard rules\n\n```yaml\n---\nname: kebab-case-name           # Must match folder. 1-64 chars, [a-z0-9-], no leading\u002Ftrailing\u002Fconsecutive hyphens.\ndescription: >                   # Folded multiline OK; ≤1024 chars; no XML angle brackets.\n  What it does. Use when [triggers]. Do NOT use for [exclusions].\nlicense: CC-BY-4.0\nmetadata:\n  version: 1.0.0\n  author: your-name-or-org\n---\n```\n\nForbidden:\n- Spaces or capitals in `name` or folder\n- `claude` \u002F `anthropic` reserved terms\n- `SKILL.MD`, `Skill.md`, `skill.md` (must be exactly `SKILL.md`)\n- XML angle brackets `\u003C` `>` in description\n- Description > 1024 characters\n\n### 3.2 Body writing principles\n\n- Imperative form (\"Run X\", \"Check Y\"). Specific > verbose.\n- Critical rules at the top, not buried in the middle.\n- 2-3 concrete examples (input → actions → result).\n- Reference files with explicit load conditions.\n- For deterministic checks, prefer scripts over prose.\n- Do not wrap prose at arbitrary column widths; let sentences flow.\n- Code blocks may wrap for readability.\n\n### 3.3 Supporting files\n\nFor each `references\u002F*.md` or `scripts\u002F*`:\n- Reference it from SKILL.md with a clear WHEN clause.\n- For files >300 lines, add a Table of Contents.\n- Scripts: see [scripts-guide](references\u002Fscripts-guide.md) for PEP 723, agentic design, --help, exit codes.\n\n**Exit:** SKILL.md drafted, all hard rules satisfied, supporting files written.\n\n---\n\n## Phase 4: VALIDATE\n\n**Goal:** Catch structural and spec errors before delivery.\n\n### 4.1 Run the bundled validator\n\n```bash\npython3 scripts\u002Fvalidate.py \u003Cskill-folder>\n# or, with PEP 723 inline dependencies via uv:\nuv run --no-project scripts\u002Fvalidate.py \u003Cskill-folder>\n```\n\nThe validator checks: folder kebab-case, SKILL.md exists, frontmatter YAML valid, `name` rules, `description` rules, body line count, examples present, references linked, no XML in description.\n\nPass = exit code 0. Warnings are allowed. Errors are not.\n\n### 4.2 Spec re-check\n\nConfirm against the full format spec (see [references\u002Fspec.md](references\u002Fspec.md)):\n- `compatibility` field if the skill needs specific environment\n- `allowed-tools` field if you want to pre-approve specific tools\n- File references are relative to the skill root, one level deep preferred\n\n### 4.3 Trigger sanity\n\nMentally run 5-10 short prompts. Should trigger; should NOT trigger (near-misses). If the description is too narrow, broaden it; if too broad, add exclusions. Save these prompts — they seed the eval in OPTIMIZE.\n\nFor the full eval-driven optimization, proceed to OPTIMIZE.\n\n**Exit:** Validator PASS + spec check + trigger sanity.\n\n---\n\n## Phase 5: OPTIMIZE\n\n**Goal:** Make the skill trigger reliably and prove it improves output quality.\n\nTwo eval loops, run in this order:\n\n### 5.1 Description triggering eval\n\nOpen [description-optimization](references\u002Fdescription-optimization.md).\n\nWrite ~20 eval queries (8-10 should-trigger, 8-10 should-NOT-trigger), emphasizing near-misses. Split 60\u002F40 into `train_queries.json` and `validation_queries.json`. Run the trigger script 3x per query. Iterate on the description (using only the train set) until stable. Select the best iteration by validation pass rate.\n\n### 5.2 Output-quality eval (optional, recommended for workflow skills)\n\nOpen [output-evaluation](references\u002Foutput-evaluation.md).\n\nBuild `evals\u002Fevals.json` with 2-3 test cases. Run each twice (with-skill, without-skill) into `iteration-1\u002F`. Add assertions, grade, aggregate to `benchmark.json`. Iterate the skill body based on failed assertions and human review. Stop when pass rate plateaus or feedback is empty.\n\n**Exit:** Description eval pass rate ≥0.5 on validation set; output eval (if run) shows skill adds value.\n\n---\n\n## Phase 6: DELIVER\n\n**Goal:** Install + register.\n\n### 6.1 Local install\n\nDrop the skill into the right skills directory for the target agent:\n\n| Agent | Path |\n|---|---|\n| VS Code + Copilot | `.agents\u002Fskills\u002F` |\n| Claude Code | `.claude\u002Fskills\u002F` |\n| Cursor | `.cursor\u002Fskills\u002F` |\n| OpenCode | `.opencode\u002Fskill\u002F` |\n\nOptionally add a slash command to the agent's command file (e.g. `\u002Fforge`) so the trigger works without typing a long phrase.\n\n### 6.2 For catalog publication\n\n1. Add a row to the catalog's index (package README) with version, description, main trigger, and link to the skill's README.\n2. Create a release note in whatever format the catalog uses (changeset, changelog entry, release tag, etc.).\n3. Bump the skill's own `metadata.version` to match the release.\n4. Publish through whatever distribution mechanism the catalog uses (npm package, git tag, registry upload).\n\n### 6.3 For distribution elsewhere\n\nKeep the skill self-contained. Make sure the skill folder is at a path the target agent scans. Include a top-level human-facing README alongside the skill folder (not inside it).\n\n**Exit:** Skill installed, registered, and a test prompt verifies triggering in the target agent.\n\n---\n\n## Bundled resources\n\n- **[references\u002Fspec.md](references\u002Fspec.md)** — the open SKILL.md format reference.\n- **[references\u002Fauthoring-patterns.md](references\u002Fauthoring-patterns.md)** — best practices: gotchas, templates, checklists, validation loops, plan-validate-execute, calibrating control.\n- **[references\u002Fdescription-optimization.md](references\u002Fdescription-optimization.md)** — trigger eval methodology with train\u002Fval split.\n- **[references\u002Foutput-evaluation.md](references\u002Foutput-evaluation.md)** — output-quality eval: with-skill vs without-skill, assertions, benchmark.\n- **[references\u002Fscripts-guide.md](references\u002Fscripts-guide.md)** — using scripts in skills: PEP 723, agentic script design, --help, exit codes.\n- **[scripts\u002Fvalidate.py](scripts\u002Fvalidate.py)** — stdlib-only validator; `--json-out` for agentic consumption.\n- **[assets\u002FSKILL.template.md](assets\u002FSKILL.template.md)** — blank template for new skills.\n\n## Anti-patterns to avoid\n\n- ❌ Vague description: \"Helps with documents.\"\n- ❌ Verbose body the agent will skim.\n- ❌ No examples.\n- ❌ Wrong `SKILL.md` casing.\n- ❌ Spaces or capitals in folder or name.\n- ❌ `claude` or `anthropic` in the name.\n- ❌ XML angle brackets in description.\n- ❌ SKILL.md > 500 lines (move detail to references\u002F).\n- ❌ Description > 1024 chars.\n- ❌ Claiming to be the only skill loaded (skills compose).\n- ❌ Re-explaining what the agent already knows (PDFs, HTTP, etc.).\n- ❌ Referencing internal projects, packages, or repos by name in the skill body.\n",{"name":5,"license":14,"metadata":51,"description":53},{"version":52},"1.0.0","Designs, authors, validates, and optimizes new Agent Skills from scratch. Use when creating a new skill, packaging a workflow as a skill, improving a skill's triggering accuracy, or evaluating whether a skill improves output quality. Covers the full lifecycle: DISCOVER, DESIGN, AUTHOR, VALIDATE, OPTIMIZE, DELIVER. Produces skills that follow the open SKILL.md format (YAML frontmatter + Markdown body, progressive disclosure, validation loops). EN triggers: \u002Fforge, create a skill, write a skill, build a skill, new skill, author a skill, package this as a skill, validate a skill, improve skill triggering, optimize skill description, evaluate skill quality. PT triggers: \u002Fforge, criar uma skill, escrever uma skill, construir uma skill, nova skill, empacotar como skill, validar skill, melhorar disparo, otimizar descrição de skill. Do NOT use for: discovering which existing skill to apply (use a dispatch\u002Fmeta skill), planning feature implementation, or designing software architecture.\n","main","33b040feb49e126600394272a8e321a3cd279e8b","packages\u002Fspells\u002Fskills\u002Fskill-forge",0,"https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum","https:\u002F\u002Fgithub.com\u002Frunecraftai","2026-07-29 22:08:52",{"source":8,"skillName":5,"commands":62,"prompt":72,"zipAvailable":73},[63,66,69],{"tool":64,"command":65},"npx","npx skills add runecraftai\u002Farcanum --skill skill-forge",{"tool":67,"command":68},"bunx","bunx skills add runecraftai\u002Farcanum --skill skill-forge",{"tool":70,"command":71},"pnpm","pnpm dlx skills add runecraftai\u002Farcanum --skill skill-forge","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fskill-forge\u002FSKILL.md",true,[75,81,87,91,95,99,103,107],{"path":76,"resourceType":77,"resourceTypeDesc":78,"sizeBytes":79,"sourceHtmlUrl":80},"assets\u002FSKILL.template.md","ASSET","资产文件",1462,"https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fskill-forge\u002Fassets\u002FSKILL.template.md",{"path":82,"resourceType":83,"resourceTypeDesc":84,"sizeBytes":85,"sourceHtmlUrl":86},"README.md","REFERENCE","参考文档",6458,"https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fskill-forge\u002FREADME.md",{"path":88,"resourceType":83,"resourceTypeDesc":84,"sizeBytes":89,"sourceHtmlUrl":90},"references\u002Fauthoring-patterns.md",10838,"https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fskill-forge\u002Freferences\u002Fauthoring-patterns.md",{"path":92,"resourceType":83,"resourceTypeDesc":84,"sizeBytes":93,"sourceHtmlUrl":94},"references\u002Fdescription-optimization.md",9134,"https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fskill-forge\u002Freferences\u002Fdescription-optimization.md",{"path":96,"resourceType":83,"resourceTypeDesc":84,"sizeBytes":97,"sourceHtmlUrl":98},"references\u002Foutput-evaluation.md",11710,"https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fskill-forge\u002Freferences\u002Foutput-evaluation.md",{"path":100,"resourceType":83,"resourceTypeDesc":84,"sizeBytes":101,"sourceHtmlUrl":102},"references\u002Fscripts-guide.md",8440,"https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fskill-forge\u002Freferences\u002Fscripts-guide.md",{"path":104,"resourceType":83,"resourceTypeDesc":84,"sizeBytes":105,"sourceHtmlUrl":106},"references\u002Fspec.md",5626,"https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fskill-forge\u002Freferences\u002Fspec.md",{"path":108,"resourceType":109,"resourceTypeDesc":110,"sizeBytes":111,"sourceHtmlUrl":112},"scripts\u002Fvalidate.py","SCRIPT","脚本",19140,"https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fskill-forge\u002Fscripts\u002Fvalidate.py",[114,138,163,188,207,230],{"skillNo":115,"name":116,"nameZh":6,"description":117,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":118,"categories":119,"keywords":121,"publishedAt":46},"01M71HGTN9FRQ3WEZMVVX7C3VZ","using-agent-skills","Discovers and invokes the right @runecraft\u002Fspells skill for the current task. Use when starting a session, when a request is ambiguous about which workflow to apply, or when you need to pick among the 13-skill Arcanum catalog. EN triggers: \u002Fskill, which skill, pick a skill, what skill should I use, skill discovery, meta skill. PT triggers: qual skill, qual habilidade, descobrir skill, qual fluxo usar. Do NOT use for: actual implementation work (route to the specific skill instead), unrelated standalone tasks outside the Arcanum catalog, or runtime-specific configuration (e.g., opencode.json \u002F Claude Code plugin paths).","https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fusing-agent-skills\u002FSKILL.md",[120],{"categoryNo":23,"categoryName":24,"taxonomyType":20,"taxonomyTypeDesc":21},[122,123,44,41,124,39,125,126,127,128,129,130,131,132,133,134,135,136,30,137],"using agent skills","standalone","pick","when","workflow","you","code","tasks","skill discovery","route","work","unrelated","opencode","request","paths","ambiguous",{"skillNo":139,"name":140,"nameZh":6,"description":141,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":142,"categories":143,"keywords":144,"publishedAt":46},"01M71HGPD6N9JH2G8EVBXJR77X","typescript-patterns","TypeScript best practices and patterns for type-safe, maintainable code. Covers discriminated unions, generics, const assertions, type narrowing, and common anti-patterns. Extends (does not override) project tsconfig. EN triggers: \u002Ftypescript, TypeScript patterns, TS best practices, type safety, discriminated unions, generics, type narrowing. PT triggers: \u002Ftypescript, padrões TypeScript, boas práticas TS, segurança de tipos, unions discriminadas, genéricos. Do NOT use for: JavaScript-only projects, runtime performance, or replacing project tsconfig rules.","https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Ftypescript-patterns\u002FSKILL.md",[],[145,146,147,148,149,150,41,151,152,153,154,155,156,157,158,159,160,161,128,162],"typescript","patterns","typescript patterns","do","projects","for","runtime","project","performance","en","maintainable code","triggers","only","use","assertions","Rules","best practices","common",{"skillNo":164,"name":165,"nameZh":6,"description":166,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":167,"categories":168,"keywords":175,"publishedAt":46},"01M71HGJAKX4TFN0PECYPJ9S38","test-driven-development","Drives development with tests — fail first, then make them pass, then refactor. Use when implementing any logic, fixing any bug, or changing any behavior. EN triggers: \u002Ftest, write tests first, TDD, prove it, 80\u002F15\u002F5 pyramid, Beyonce Rule. PT triggers: \u002Fteste, testes primeiro, provar com teste, pirâmide 80\u002F15\u002F5. Do NOT use for: production incident triage (use \u002Fdebug), or cosmetic-only changes with no behavioral surface.","https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Ftest-driven-development\u002FSKILL.md",[169,172],{"categoryNo":170,"categoryName":171,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000007","代码审查",{"categoryNo":173,"categoryName":174,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000008","测试与质量",[176,177,178,179,41,180,181,182,157,125,32,183,184,185,186,148,150,154,156,187],"test driven development","development","test","driven","incident","implementing","write tests","bug","refactor","behavioral","make","tests",{"skillNo":189,"name":190,"nameZh":6,"description":191,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":192,"categories":193,"keywords":194,"publishedAt":46},"01M71HGC1H4JQFJWH613XTJC59","spec-driven","Feature planning and implementation with 4 adaptive phases — Specify, Design, Tasks, Execute. Auto-sizes depth by complexity. Creates atomic tasks with verification criteria, atomic git commits, and requirement traceability. Features an independent Verifier (author != verifier, evidence-or-zero), persistent decision log (STATE.md), and test-coverage-matrix-driven tests, plus a self-improving lessons layer that turns verification failures into reusable project-local guidance. Stack-agnostic. Use when planning features (requirements, design, task breakdown), implementing with verification and atomic commits, or validating\u002Fverifying an implementation against a spec. Triggers (PT\u002FEN): \"specify feature\", \"vamos especificar\", \"discutir feature\", \"discutir este caso\", \"design\", \"design da feature\", \"tarefas\", \"quebrar em tarefas\", \"implementar\", \"build\", \"validar\", \"verify work\", \"UAT\", \"validar implementação\", \"record decision\", \"pause work\", \"pausar trabalho\", \"resume work\", \"retomar trabalho\". Do NOT use for architecture decomposition analysis (use architecture skills) or technical design docs (treat as Complex scope: discuss gray areas + architecture-focused design).","https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fspec-driven\u002FSKILL.md",[],[195,196,179,197,198,199,200,41,152,201,181,202,36,203,204,205,125,206,129,35],"spec driven","spec","coverage","complexity","task breakdown","scope","decision","local","self-improving","auto","atomic commits","analysis",{"skillNo":208,"name":209,"nameZh":6,"description":210,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":211,"categories":212,"keywords":213,"publishedAt":46},"01M71HG1RAQY3GXX0Q5TMW0XD6","shipping-and-launch","Prepares production launches with a pre-launch checklist, staged rollout, monitoring, and a rollback strategy — plus feature flag lifecycle (off → internal → beta → GA). Use when deploying a feature to production for the first time, releasing a significant change, or planning a rollout. EN triggers: \u002Fship, pre-launch checklist, staged rollout, feature flag, rollback plan, monitoring. PT triggers: \u002Fshippar, checklist pré-lançamento, rollout gradual, feature flag, plano de rollback. Do NOT use for: pre-production development work, or one-off internal tools with no user exposure.","https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fshipping-and-launch\u002FSKILL.md",[],[214,215,216,41,217,218,125,219,220,221,222,132,223,224,225,226,33,227,228,229],"shipping and launch","shipping","launch","time","beta","rollout","rollback plan","monitoring","rollback strategy","plan","change","exposure","plano","ship","Tools","pre-launch",{"skillNo":231,"name":232,"nameZh":6,"description":233,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":234,"categories":235,"keywords":241,"publishedAt":46},"01M71HFXGHP092ZKFR762M9394","security-and-hardening","Hardens code against vulnerabilities using the OWASP Top 10 and a three-tier boundary system (input, processing, output). Use when handling user input, authentication, data storage, or external integrations. EN triggers: \u002Fsecurity, security audit, OWASP, three-tier boundary, least privilege, input validation. PT triggers: \u002Fsegurança, auditoria de segurança, OWASP, privilégio mínimo, validar entrada. Do NOT use for: non-security performance work (use \u002Freview's performance axis), or dependency updates that have no security exposure.","https:\u002F\u002Fgithub.com\u002Frunecraftai\u002Farcanum\u002Fblob\u002Fmain\u002Fpackages\u002Fspells\u002Fskills\u002Fsecurity-and-hardening\u002FSKILL.md",[236],{"categoryNo":237,"categoryName":238,"taxonomyType":239,"taxonomyTypeDesc":240},"01JBSEEDCAT000000000000033","数据分析师","PROFESSION","适用职业",[242,243,244,41,245,225,33,246,148,247,150,248,249,156,158,250,251,252,253,254],"security","hardening","security and hardening","authentication","input","three","system","least privilege","privilege","validation","vulnerabilities","user","processing"]