All skills
simota avatar

/sigil

@35ffd55
by shingo imotasimota/agent-skills85 stars
15

Designing a repository's project-local operating layer and generating its skills, recipes, workflows, and routing map. Not for global ecosystem agents (Architect) or runtime execution (Nexus).

Use this Skill: https://skilld.dev/gh/simota/agent-skills/sigil

This session only. Nothing lands on disk.

referenceskill-templates.md

≈1.3k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Skill Templates

Purpose: load this during CRAFT to choose Micro vs Full, preserve required structure, and fill templates with project-specific conventions.

Contents

  1. Frontmatter rules
  2. Type selection
  3. Micro template
  4. Full template
  5. Authoring guardrails
  6. File and size rules

Frontmatter Rules

Every generated skill MUST include YAML frontmatter:

---
name: skill-name
description: [one Japanese sentence]
---

Ecosystem-specific constraints (additive to the official spec — see official-skill-guide.md § 2 YAML Frontmatter Full Specification for the canonical kebab-case / 1024-char / XML / reserved-name rules and the rationale behind them):

  • name: unique within the project, normally 2-4 words
  • description: one Japanese sentence, concise, routing-friendly (project guardrail — official limit is 1024 chars)
  • Add optional fields only when the generated skill truly needs them

Type Selection

Use this type When Expected size
Micro Single task, clear outcome, 0-2 decision points 10-80 lines
Full Multi-step process, 3+ decision points, significant domain knowledge, or multiple variants 100-400 lines

Promote to Full when rollback guidance, multiple patterns, or supporting references are required for safe execution.

Micro Template

Use English headings by default, but match the project's established skill language if existing skills use another convention.

---
name: [skill-name]
description: [one Japanese sentence]
---

# [Skill Title]

## Purpose
[Explain when and why to use this skill.]

## Steps

1. [Step 1]
2. [Step 2]
3. [Step 3]

## Template

\```[lang]
[Project-specific starter template]
\```

## Conventions
- [Naming rule]
- [File placement rule]
- [Testing or error-handling rule]

Micro minimum:

  • Purpose: 1-2 sentences
  • Steps: 3+
  • Include at least one of Template or Conventions

Full Template

---
name: [skill-name]
description: [one Japanese sentence]
---

# [Skill Title]

## Purpose
[Explain scope, prerequisites, and expected outcome.]

## Design Principles
- [Project-specific principle 1]
- [Project-specific principle 2]
- [Project-specific principle 3]

## Workflow

### Step 1: [Phase Name]
[Actionable instructions]

### Step 2: [Phase Name]
[Actionable instructions]

### Step 3: [Phase Name]
[Actionable instructions]

## Templates

### Pattern A: [Simple case]

\```[lang]
[Template code]
\```

### Pattern B: [Complex case]

\```[lang]
[Template code]
\```

## Error Handling
[Project-specific failure and recovery rules]

## Testing
[Framework-specific test expectations]

## Checklist
- [ ] [Check 1]
- [ ] [Check 2]
- [ ] [Check 3]

Full minimum:

  • Purpose: 3+ sentences including prerequisites
  • Workflow: 3+ phases
  • Templates: 2+ patterns or variations
  • Explicit Error Handling, Testing, and Checklist

Authoring Guardrails

Variable Substitution

Variable Source Example
[ProjectName] manifest name my-app
[Framework] framework detection Next.js
[FrameworkVersion] manifest dependencies 14.1.0
[ComponentDir] directory scan src/components
[TestDir] test layout detection __tests__
[TestFramework] dependencies vitest
[CSSApproach] config or deps tailwind
[PackageManager] lock files pnpm
[ImportAlias] tsconfig.json @/

Rules:

  1. Replace every [Variable] before install.
  2. If the value is unknown, use the safest framework default and add <!-- TODO: verify -->.
  3. If multiple values are valid, use conditional branches instead of guessing.
  4. Never ship raw placeholders in installed skills.

Structural Guardrails

  • Every step must be executable in the target project.
  • All code blocks must have language tags.
  • Use the project's actual paths and dependency set.
  • Do not hardcode developer-specific machine paths.
  • Keep references one hop away from SKILL.md.
  • For Claude Code-specific fields and sandbox choices, cross-check claude-code-skills-api.md.
  • For final scoring, use validation-rules.md.

File and Size Rules

Required Layout

  • .claude/skills/[skill-name]/SKILL.md
  • .agents/skills/[skill-name]/SKILL.md
  • Full Skills may add:
    • .claude/skills/[skill-name]/reference/[topic].md
    • .agents/skills/[skill-name]/reference/[topic].md

Naming Rules

Project-specific layout (canonical naming and file-structure rules — directory kebab-case, exact SKILL.md spelling, no README.md inside skill folders — live in official-skill-guide.md § 5 File Structure Requirements):

  • Reference files: reference/[topic].md (one hop from SKILL.md)
  • Skills are installed to both .claude/skills/ and .agents/skills/ (Sigil-specific dual-install)

Size Guidelines

Type Lines Sections Templates
Micro 10-80 3-5 0-1
Full 100-400 6-10 2-5

Source: SKILL.md on GitHub

1 warning13d5 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    The skill is a project-local development tool designed to analyze codebase context and generate AI agent skills. It uses shell commands and local file access to detect tech stacks and conventions. While it includes safety guardrails against prompt injection and credential exposure, it possesses an attack surface for indirect prompt injection, where a malicious repository could influence the generation of executable skill instructions or configurations.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    2/12 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 35ffd55. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 3 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/sigil