skill-architect

name	skill-architect
description	Design, create, audit, and improve Claude Agent Skills with expert-level progressive disclosure. Use when building new skills, reviewing existing skills, debugging activation failures, encoding Use when this capability is needed.
metadata	{"author":"curiositech"}

Skill Architect: The Authoritative Meta-Skill

The unified authority for creating expert-level Agent Skills. Encodes the knowledge that separates a skill that merely exists from one that activates precisely, teaches efficiently, and makes users productive immediately.

Philosophy

Great skills are progressive disclosure machines. They encode real domain expertise (shibboleths), not surface instructions. They follow a three-layer architecture: lightweight metadata for discovery, lean SKILL.md for core process, and reference files for deep dives loaded only on demand.

---

When to Use This Skill

✅ Use for:

• Creating new skills from scratch or from existing expertise
• Auditing/reviewing skills for quality, activation, and progressive disclosure
• Improving activation rates and reducing false positives
• Encoding domain expertise (shibboleths, anti-patterns, temporal knowledge)
• Designing skills that subagents consume effectively
• Building self-contained tools (scripts, MCPs, subagents)
• Debugging why skills don't activate or activate incorrectly

❌ NOT for:

• General Claude Code features (slash commands, MCP server implementation)
• Non-skill coding advice or code review
• Debugging runtime errors (use domain-specific skills)
• Template generation without real domain expertise to encode

---

Quick Wins (Immediate Improvements)

For existing skills, apply in priority order:

1. Tighten description → Follow [What] [When to use]. NOT for [Exclusions] formula
2. Check line count → SKILL.md must be <500 lines; move depth to /references
3. Add NOT clause → Prevent false activation with explicit exclusions
4. Add 1-2 anti-patterns → Use shibboleth template (Novice/Expert/Timeline)
5. Remove dead files → Delete unreferenced files in scripts/ and references/ (no phantoms)
6. Test activation → Write 5 queries that should trigger and 5 that shouldn't

---

Progressive Disclosure Architecture

Skills use three-layer loading. The runtime scans metadata at startup, loads SKILL.md on activation, and pulls reference files only when the agent decides it needs them.

Layer	Content	Size	Loading
1. Metadata	name + description in frontmatter	~100 tokens	Always in context (catalog scan)
2. SKILL.md	Core process, decision trees, brief anti-patterns	<5k tokens	On skill activation
3. References	Deep dives, examples, templates, specs	Unlimited	On-demand, per-file, only when relevant

Critical rules:

• Keep SKILL.md under 500 lines. Move depth to /references.
• Reference files are NOT auto-loaded. Only SKILL.md enters context on activation.
• In SKILL.md, list each reference file with a 1-line description of when to consult it. This teaches the agent what's available without loading it.
• Never instruct "read all reference files before starting." Instead: "Read only the files relevant to the current step."
• If a reference file is large, the agent should skim headings first, then drill into the relevant section.

---

Frontmatter Rules

Required Fields

Key	Purpose	Constraints
name	Lowercase-hyphenated identifier	Max 64 chars, a-z/0-9/hyphens only, no "anthropic" or "claude", no XML tags
description	Activation trigger: [What] [When to use]. NOT for [Exclusions]	Max 1024 chars, no XML tags. See Description Formula

Optional Fields

Key	Purpose	Example
allowed-tools	Comma-separated tool names (least privilege)	Read,Write,Grep
argument-hint	Hint shown in autocomplete for expected arguments	"[path] [format]"
license	License identifier	MIT
disable-model-invocation	If true, only user-triggered via /skill-name	true
user-invocable	Controls whether skill appears in UI menus	true
context	Execution context; fork runs skill in isolated subagent	fork
agent	Which subagent type when context: fork	code-reviewer
model	Override model when skill is active	sonnet
hooks	Hooks scoped to this skill's lifecycle	See hooks reference
metadata	Arbitrary key-value map for tooling/dashboards	author: your-org
dependencies	Required packages for scripts	python>=3.8, pandas>=1.5.0
bundled-resources	Declared resource files	(list of paths)
distribution	Distribution method	npm, pip, zip

Note: The Skills API uses hyphenated variants (context-fork, model-override). Claude Code uses the names shown above.

Custom Keys (Safe to Use)

Custom keys like category, tags, version are ignored by Claude Code but safe to include for your own tooling (gallery websites, documentation generators, dashboards). They don't conflict with runtime parsing.

Invalid Keys (Confusingly Similar to Valid Ones)

# ❌ These look like valid keys but aren't — use the correct alternatives
tools: Read,Write           # Use 'allowed-tools' instead
integrates_with: [...]      # Use SKILL.md body text instead
outputs: [...]              # Use SKILL.md Output Format section instead
python_dependencies: [...]  # Use 'dependencies' in frontmatter instead

---

Platform Constraints

Constraint	Limit
Name length	64 characters
Name format	Lowercase letters, numbers, hyphens only
Reserved words in name	"anthropic", "claude" prohibited
Description length	1024 characters
Total skill upload size	8MB (API and claude.ai)
Skills per API request	8 maximum
XML tags in name/description	Prohibited

Cross-surface behavior: Skills do NOT sync across Claude.ai, Claude API, and Claude Code. Each surface requires separate upload/management. Maintain source files in Git as single source of truth.

Recall limits: Too many active skills degrades Claude's selection accuracy. Test coexistence when adding new skills — verify the new skill doesn't steal triggers from existing ones.

---

Description Formula

Pattern: [What it does] [When to use — be slightly pushy]. NOT for [Exclusions].

The description is the most important line for activation. Claude's runtime scans descriptions to decide which skill to load. A weak description means zero activations or constant false positives.

How activation works: Claude evaluates descriptions semantically, not via keyword matching. It reasons about whether your description covers the user's intent. This means specific, context-rich descriptions outperform keyword lists. Claude also tends to undertrigger — it errs toward NOT activating skills. Combat this by making descriptions slightly pushy: explain when the skill should be used, even if it seems obvious.

Problem	Bad	Good
Too vague	"Helps with images"	"CLIP semantic search for image-text matching and zero-shot classification. NOT for counting, spatial reasoning, or generation."
No exclusions	"Reviews code changes"	"Reviews TypeScript/React diffs and PRs for correctness. NOT for writing new features."
Mini-manual	"Researches, then outlines, then drafts..."	"Structured research producing 1-3 page synthesis reports. NOT for quick factual questions."
Catch-all	"Helps with product management"	"Writes and refines product requirement documents (PRDs). NOT for strategy decks."
Name mismatch	name: db-migration / desc: "writes marketing emails"	name: db-migration / desc: "Plans database schema migrations with rollback strategies."

Full guide with more examples: See references/description-guide.md

---

SKILL.md Template

---
name: your-skill-name
description: [What it does] [When to use — be slightly pushy]. NOT for [Exclusions].
allowed-tools: Read,Write
---

# Skill Name
[One sentence purpose]

## When to Use
✅ Use for: [A, B, C with specific trigger keywords]
❌ NOT for: [D, E, F — explicit boundaries]

## Core Process
[Mermaid diagrams — 23 types available. See visual-artifacts.md for full catalog]

## Anti-Patterns
### [Pattern Name]
**Novice**: [Wrong assumption]
**Expert**: [Why it's wrong + correct approach]
**Timeline**: [When this changed, if temporal]

## References
- `references/guide.md` — Consult when [specific situation]
- `references/examples.md` — Consult for [worked examples of X]

---

The 6-Step Skill Creation Process

flowchart LR
  S1[1. Gather Examples] --> S2[2. Plan Contents]
  S2 --> S3[3. Initialize]
  S3 --> S4[4. Write Skill]
  S4 --> S5[5. Validate]
  S5 --> S6{Errors?}
  S6 -->|Yes| S4
  S6 -->|No| S7[6. Ship & Iterate]

Step 1: Gather Concrete Examples

Collect 3-5 real queries that should trigger this skill, and 3-5 that should NOT.

Step 2: Plan Reusable Contents

For each example, identify what scripts, references, or assets would prevent re-work. Also identify shibboleths: domain algorithms, temporal knowledge, framework evolution, common pitfalls.

Step 3: Initialize

scripts/init_skill.py <skill-name> --path <output-directory>

For existing skills, skip to Step 4.

Step 4: Write the Skill

Order of implementation:

1. Scripts first (scripts/) — Working code, not templates
2. References next (references/) — Domain knowledge, schemas, guides
3. SKILL.md last — Core process, anti-patterns, reference index

Write in imperative form: "To accomplish X, do Y" not "You should do X."

Answer these questions in SKILL.md:

1. Purpose: What is this skill for? (1-2 sentences)
2. Activation: What triggers it? What shouldn't?
3. Process: Use Mermaid diagrams (23 types) — flowcharts for decisions, sequence for protocols, state for lifecycles, etc.
4. Anti-patterns: What do novices get wrong?
5. Visual artifacts: Render workflows, architectures, timelines as Mermaid diagrams (see references/visual-artifacts.md)
6. References: What files exist and when to consult them?

Step 5: Validate

python scripts/validate_skill.py <path>
python scripts/check_self_contained.py <path>

Fix ERRORS → WARNINGS → SUGGESTIONS.

Step 6: Iterate

After real-world use: notice struggles, improve SKILL.md and resources, update CHANGELOG.md.

---

Designing Skills for Subagent Consumption

When skills will be loaded by subagents (not just direct user invocation), apply these patterns:

Three Skill-Loading Layers

1. Preloaded (2-5 core skills): Injected into the subagent's system context. These are its standard operating procedures — always present.
2. Dynamically selected: Subagent receives a catalog (name + 1-line description) and picks 1-3 matching skills before starting. The orchestrator can also pre-filter.
3. Execution-time: Subagent reads each skill's "When to use" section, follows numbered steps in order, respects output contracts, and runs QA checks.

How Subagents Should Use Skills

Teach the subagent to treat each skill like a mini-protocol:

• Check the "When to use / When not to use" section for applicability
• Follow numbered steps in order (adapt only if task constraints force it)
• Respect the skill's output contract (templates, JSON shapes, required sections)
• Apply QA/validation steps last
• Reference skill steps by number: "Completed step 3 of refactor-plan-skill"

Subagent Prompt Structure

The subagent's prompt should have four sections:

1. Identity: "You are the [role]. You handle [narrow domain]. If outside scope, say so."
2. Skill usage rules: "Your skills define your methods. Decide which apply, follow their workflows."
3. Task loop: Restate → Select skills → Clarify → Plan → Execute step-by-step → Validate → Return (artifacts + skills used + remaining risks).
4. Constraints: Quality bar, safety rules, tie-breaking priorities.

Full templates and orchestration patterns: See references/subagent-design.md

---

Visual Artifacts: Mermaid Diagrams & Code

Skills that include Mermaid diagrams serve two audiences at once. For humans, diagrams render as visual flowcharts, state machines, and timelines — instantly parseable. For agents, Mermaid is a text-based graph DSL — A -->|Yes| B is an explicit, unambiguous edge that's actually easier to reason about than equivalent prose. The agent reads the text; the human sees the picture. Both win.

Rule: If a skill describes a process, decision tree, architecture, state machine, timeline, or data relationship, include a Mermaid diagram. Use raw ```mermaid blocks directly in SKILL.md — not wrapped in outer markdown fences.

Most Useful Diagram Types for Skills

Skill Content	Diagram Type	Syntax
Decision trees / troubleshooting	Flowchart	flowchart TD
API/agent communication protocols	Sequence	sequenceDiagram
Lifecycle / status transitions	State	stateDiagram-v2
Temporal knowledge / evolution	Timeline	timeline
Data models / schemas	ER	erDiagram
Domain taxonomy / concept maps	Mindmap	mindmap
Priority matrices (2-axis)	Quadrant	quadrantChart
Infrastructure / cloud topology	Architecture	architecture-beta

Full catalog (all 23 types) with syntax, examples, and YAML config: See references/visual-artifacts.md

---

Encoding Shibboleths

Expert knowledge that separates novices from experts. Things LLMs get wrong due to outdated training data or cargo-culted patterns.

Shibboleth Template

### Anti-Pattern: [Name]
**Novice**: "[Wrong assumption]"
**Expert**: [Why it's wrong, with evidence]
**Timeline**: [Date]: [Old way] → [Date]: [New way]
**LLM mistake**: [Why LLMs suggest the old pattern]
**Detection**: [How to spot this in code/config]

What to Encode

• Framework evolution (React Classes → Hooks → Server Components)
• Model limitations (CLIP can't count; embedding models are task-specific)
• Tool architecture (Script → MCP graduation path)
• API versioning (ada-002 → text-embedding-3-large)
• Temporal traps (advice that was correct in 2023 but harmful in 2025)

Full catalog with case studies: See references/antipatterns.md

---

Self-Contained Tools and the Extension Taxonomy

Skills are one of seven Claude extension types: Skills (domain knowledge), Plugins (packaged bundles for distribution), MCP Servers (external APIs + auth), Scripts (local operations), Slash Commands (user-triggered skills), Hooks (lifecycle automation at 17+ event points), and Agent SDK (programmatic Claude Code access). Most skills should include scripts. MCPs are only for auth/state boundaries. Plugins are for sharing skills across teams/community.

Need	Extension Type	Key Requirement
Domain expertise / process	Skill (SKILL.md)	Decision trees, anti-patterns, output contracts
Packaging & distribution	Plugin (plugin.json)	Bundles skills + hooks + MCP + agents
External API + auth	MCP Server	Working server + setup README
Repeatable local operation	Script	Actually runs (not a template), minimal deps
Multi-step orchestration	Subagent	4-section prompt, skills, workflow
User-triggered action	Slash Command	Skill with user-invocable: true
Lifecycle automation	Hook	17+ events: PreToolUse, PostToolUse, Stop, etc.
Programmatic access	Agent SDK	npm/pip package, CI/CD pipelines

Evolution path: Skill → Skill + Scripts → Skill + MCP Server → Skill + Subagent → Plugin (for distribution). Only promote when complexity justifies it.

Full taxonomy with examples and common mistakes: See references/claude-extension-taxonomy.md Detailed tool patterns: See references/self-contained-tools.md Plugin creation and distribution: See references/plugin-architecture.md

---

Tool Permissions (Least Privilege)

Access Level	allowed-tools
Read-only	Read,Grep,Glob
File modifier	Read,Write,Edit
Build integration	Read,Write,Bash(npm:*,git:*)
⚠️ Never for untrusted	Unrestricted Bash

---

Anti-Pattern Summary

#	Anti-Pattern	Fix
1	Documentation Dump	Decision trees in SKILL.md, depth in /references
2	Missing NOT clause	Always include "NOT for X, Y, Z" in description
3	Phantom Tools	Only reference files that exist and work
4	Template Soup	Ship working code or nothing
5	Overly Permissive Tools	Least privilege: specific tool list, scoped Bash
6	Stale Temporal Knowledge	Date all advice, update quarterly
7	Catch-All Skill	Split by expertise type, not domain
8	Vague Description	Use [What] [When to use]. NOT for [Exclusions]
9	Eager Loading	Never "read all files first"; lazy-load references
10	Prose-Only Processes	Use Mermaid diagrams (23 types) — flowcharts, sequences, states, ER, timelines, etc.

Full case studies: See references/antipatterns.md

---

Validation Checklist

□ SKILL.md exists and is <500 lines
□ Frontmatter has name + description (minimum required)
□ Description follows [What][When to use] NOT [Exclusions] formula
□ Description is specific and context-rich (semantic activation, not keyword lists)
□ Name and description are aligned (not contradictory)
□ At least 1 anti-pattern with shibboleth template
□ All referenced files actually exist (no phantoms)
□ Scripts work (not templates), have clear CLI, handle errors
□ Reference files each have a 1-line purpose in SKILL.md
□ Processes/decisions/lifecycles use Mermaid diagrams (23 types), not prose
□ CHANGELOG.md tracks version history
□ If subagent-consumed: output contracts are defined

Run automated checks: python scripts/validate_skill.py <path> and python scripts/validate_mermaid.py <path>

---

Common Rejection Causes

Cause	Symptom	Fix
Missing name or description	Skill won't load	Add both to frontmatter
tools: instead of allowed-tools:	Tools silently ignored	Use allowed-tools: (hyphenated)
YAML list in allowed-tools	Parse error	Use comma-separated: Read,Write,Edit
Brackets in allowed-tools	Parse error	No [ ] — just Read,Write,Edit
Invalid keys (triggers, outputs)	Silently ignored or error	Move to SKILL.md body text
Name with spaces/uppercase	May fail matching	Lowercase-hyphenated: my-skill-name
Name doesn't match directory	Activation mismatch	Keep name = directory name
context: not fork	Ignored	Only valid value is fork
disable-model-invocation: not boolean	Ignored	Use true or false
Phantom file references	Agent wastes tool calls	Delete references or create files

Full validation: python scripts/validate_skill.py <path> catches all of these.

---

Success Metrics

Metric	Target
Correct activation	>90%
False positive rate	<5%
Token usage	<5k tokens
Time to productive	<5 min

---

Reference Files

Consult these for deep dives — they are NOT loaded by default:

File	Consult When
references/knowledge-engineering.md	KE methods for extracting expert knowledge into skills; protocol analysis, repertory grids, aha! moments
references/description-guide.md	Writing or rewriting a skill description
references/antipatterns.md	Looking for shibboleths, case studies, or temporal patterns
references/self-contained-tools.md	Adding scripts, MCP servers, or subagents to a skill
references/subagent-design.md	Designing skills for subagent consumption or orchestration
references/claude-extension-taxonomy.md	Skills vs Plugins vs MCPs vs Hooks vs Agent SDK — the 7-type taxonomy
references/plugin-architecture.md	Creating, packaging, and distributing plugins via marketplaces
references/visual-artifacts.md	Adding Mermaid diagrams: all 23 types, YAML config, best practices
references/mcp-template.md	Building an MCP server for a skill
references/subagent-template.md	Defining subagent prompts and multi-agent pipelines
references/scoring-rubric.md	Quantitative skill evaluation (0-10 scoring criteria)
references/skill-composition.md	Cross-skill dependencies and composition patterns
references/skill-lifecycle.md	Maintenance, versioning, and deprecation guidance
references/activation-debugging.md	Diagnosing why skills don't activate or false-positive; systematic debugging steps
agents/cross-evaluator.md	Template for cross-evaluating skills — inject source expertise, evaluate target, produce improved version

---

Source: curiositech/some_claude_skills — distributed by TomeVault.