Agent skill

workflows-multicam-qara-multicam-qara

Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/testing/workflows-multicam-qara-multicam-qara

SKILL.md

Validate Skill Workflow

Purpose: Audit existing skill for compliance with canonical PAI architectural standards

When to Use:

  • User says "validate skill", "check skill compliance", "audit skill structure"
  • Before deploying a skill to production
  • After creating or updating a skill
  • When reviewing skill quality
  • As part of skill maintenance

Prerequisites:

  • Target skill exists in ${PAI_DIR}/skills/
  • Access to skill-structure.md
  • Understanding of validation criteria

Workflow Steps

Step 1: Read Canonical Architecture

REQUIRED FIRST STEP: Read the source of truth:

bash
${PAI_DIR}/skills/CORE/skill-structure.md

What to extract:

  • The 3 archetypes (Minimal, Standard, Complex)
  • Mandatory structural requirements
  • Workflow routing rules
  • Naming conventions
  • Quality checklist

Output: Validation criteria loaded


Step 2: Identify Target Skill

If user specifies skill name:

bash
${PAI_DIR}/skills/[skill-name]/

If user says "validate this skill":

  • Check current working directory
  • If in skill directory (contains SKILL.md), use that
  • Otherwise ask user to specify

Verify skill exists:

bash
test -f ${PAI_DIR}/skills/[skill-name]/SKILL.md && echo "✅ Skill found" || echo "❌ Skill not found"

Output: Target skill identified


Step 3: Structural Validation

Check directory structure:

bash
# List all files
tree ${PAI_DIR}/skills/[skill-name]/

# Or if tree not available:
find ${PAI_DIR}/skills/[skill-name]/ -type f -o -type d

Validate archetype compliance:

For Minimal Skill (0-3 workflows):

  • SKILL.md present
  • workflows/ OR assets/ directory present
  • No unnecessary directories

For Standard Skill (3-15 workflows):

  • SKILL.md present
  • workflows/ directory present
  • Optional: documentation/, references/, tools/
  • No unnecessary complexity

For Complex Skill (15+ workflows):

  • SKILL.md present
  • workflows/ directory (nested structure)
  • documentation/ directory (CORE is flat exception)
  • Optional: CONSTITUTION.md, METHODOLOGY.md
  • Optional: references/, state/, tools/, testing/

Count workflows:

bash
find ${PAI_DIR}/skills/[skill-name]/workflows/ -name "*.md" -type f | wc -l

Determine expected archetype:

  • 0-3 workflows → Should be Minimal
  • 3-15 workflows → Should be Standard
  • 15+ workflows → Should be Complex

Check archetype match:

  • ❓ Does structure match workflow count?
  • ❓ Is skill over-engineered (Complex structure with 5 workflows)?
  • ❓ Is skill under-engineered (Minimal structure with 20 workflows)?

Naming conventions:

bash
# Check SKILL.md is uppercase
test -f ${PAI_DIR}/skills/[skill-name]/SKILL.md && echo "✅ SKILL.md correct" || echo "❌ Should be SKILL.md"

# Check workflow naming (should be kebab-case)
find ${PAI_DIR}/skills/[skill-name]/workflows/ -name "*.md" -exec basename {} \; | grep -v '^[a-z][a-z0-9-]*\.md$' && echo "❌ Non-kebab-case workflows found" || echo "✅ Workflows kebab-case"

Score: [X/10]

Output: Structural validation results


Step 4: Routing Validation

Read SKILL.md:

bash
${PAI_DIR}/skills/[skill-name]/SKILL.md

Check YAML frontmatter:

  • name: field present
  • description: field present
  • Description includes USE WHEN triggers
  • Description has 5-10 trigger phrases

Check Workflow Routing section:

  • "Workflow Routing (SYSTEM PROMPT)" section present
  • Section is FIRST (immediately after YAML frontmatter)
  • NOT buried in middle or end of file

Count routed workflows:

bash
# Manual inspection - count "When user requests" blocks
grep -c "When user requests" ${PAI_DIR}/skills/[skill-name]/SKILL.md

Compare to actual workflows:

bash
# Count actual workflow files
find ${PAI_DIR}/skills/[skill-name]/workflows/ -name "*.md" -type f | wc -l

Validation:

  • ❓ Are all workflows routed? (route count = file count)
  • ❓ Are orphan workflows present? (files not routed)
  • ❓ Are dead routes present? (routes to non-existent files)

Check routing quality: For each route:

  • Examples provided (3-5 user phrases)
  • File path is absolute (${PAI_DIR}/skills/...)
  • EXECUTE description provided
  • Examples are semantic (natural language), not formulaic

Example of GOOD routing:

markdown
**When user requests person research:**
Examples: "do OSINT on [person]", "research [person]", "background check on [person]", "who is [person]", "investigate this person"
→ **READ:** ${PAI_DIR}/skills/security-OSINT/workflows/people/lookup.md
→ **EXECUTE:** Complete person OSINT workflow

Example of BAD routing:

markdown
**When user requests person research:**
→ READ: workflows/people/lookup.md
→ EXECUTE: Person OSINT

Score: [X/10]

Output: Routing validation results


Step 5: Activation Triggers Validation

Check "When to Activate This Skill" section:

  • Section present
  • Located after Workflow Routing section

8-Category Pattern Coverage:

Check if skill covers these categories:

  1. Core Skill Name (Noun) - ❓ Present?

    • Skill name variations
    • Abbreviations
  2. Action Verbs - ❓ Present?

    • "do [skill]", "run [skill]", "perform [skill]", etc.
  3. Modifiers (Scope/Intensity) - ❓ Present?

    • "basic [skill]", "quick [skill]", "comprehensive [skill]", etc.
  4. Prepositions (Target Connection) - ❓ Present?

    • "[skill] on [target]", "[skill] for [target]", etc.
  5. Synonyms & Alternative Phrasings - ❓ Present?

    • Industry jargon, casual vs formal
  6. Use Case Oriented - ❓ Present?

    • Why would someone use this?
    • What problem does it solve?
  7. Result-Oriented Phrasing - ❓ Present?

    • "find [thing]", "discover [thing]", "get [information]"
  8. Tool/Method Specific - ❓ Present if applicable?

    • Specific tools or techniques

Category Coverage Score: [X/8]

Comprehensiveness check:

  • Covers at least 5/8 categories
  • Includes casual phrasing ("just do X", "super basic X")
  • Includes natural variations (not just formulaic)
  • Passes "read aloud" test (sounds like real requests)

Score: [X/10]

Output: Activation trigger validation results


Step 6: Documentation Validation

Check file linkage:

List all .md files in skill:

bash
find ${PAI_DIR}/skills/[skill-name]/ -name "*.md" -type f

For each file (excluding SKILL.md):

  • ❓ Is it referenced in SKILL.md main body?
  • ❓ Is purpose explained?
  • ❓ Is when-to-use guidance provided?

Check for orphan files:

  • Files in skill directory not linked from SKILL.md

Check for broken links:

  • SKILL.md references non-existent files

Documentation quality:

  • Each workflow has clear purpose
  • Each workflow has when-to-use guidance
  • Examples provided
  • Related workflows linked

Extended context sections:

  • Core capabilities explained
  • Integration points documented
  • Configuration documented (if applicable)
  • Examples provided

Score: [X/10]

Output: Documentation validation results


Step 7: Integration Validation

Check for CORE duplication:

  • Skill doesn't duplicate CORE context
  • References CORE instead of copying
  • Self-contained but inherits CORE

Check agent compatibility:

  • Can be invoked via Skill tool
  • Compatible with agent workflows
  • No blocking dependencies

Score: [X/10]

Output: Integration validation results


Step 8: Quality Validation

Progressive disclosure check:

  • SKILL.md is the hub (routing + overview)
  • Workflows contain detailed steps
  • Documentation provides deep context
  • Not all info crammed in SKILL.md

Naming convention compliance:

  • SKILL.md is UPPERCASE
  • Root docs are UPPERCASE (METHODOLOGY, CONSTITUTION)
  • Workflows are kebab-case
  • Directories follow conventions (workflows, tools, documentation)

Template compliance:

  • Follows canonical SKILL.md template
  • Workflow files follow standard template
  • Consistent structure across files

Examples present:

  • SKILL.md has usage examples
  • Workflows have input/output examples
  • Clear demonstration of capabilities

Score: [X/10]

Output: Quality validation results


Step 9: Generate Validation Report

Compile all results:

markdown
# Skill Validation Report: [skill-name]

**Date:** [YYYY-MM-DD]
**Validator:** system-create-skill
**Overall Status:** [PASS / FAIL / NEEDS IMPROVEMENT]

---

## Executive Summary

**Overall Score:** [X/70] ([percentage]%)
**Archetype:** [Minimal/Standard/Complex]
**Compliance Status:** [Compliant/Non-Compliant]

**Critical Issues:** [N]
**Major Issues:** [N]
**Minor Issues:** [N]

---

## Validation Results

### 1. Structural Validation: [X/10]

**Archetype Compliance:**
- Expected: [archetype based on workflow count]
- Actual: [archetype based on structure]
- Match: [Yes/No]

**Directory Structure:**
✅ [Compliant elements]
❌ [Non-compliant elements]
⚠️ [Issues needing attention]

**Naming Conventions:**
✅ [Correct naming]
❌ [Incorrect naming]

---

### 2. Routing Validation: [X/10]

**Workflow Routing Section:**
- Present: [Yes/No]
- Location: [First/Middle/End/Missing]
- ✅/❌ Positioning correct

**Workflow Coverage:**
- Total workflows: [N]
- Routed workflows: [N]
- Orphan workflows: [N]
- Dead routes: [N]

**Routing Quality:**
✅ [Good routing examples]
❌ [Poor routing examples]

---

### 3. Activation Triggers: [X/10]

**8-Category Pattern Coverage: [X/8]**
1. Core Skill Name: [✅/❌]
2. Action Verbs: [✅/❌]
3. Modifiers: [✅/❌]
4. Prepositions: [✅/❌]
5. Synonyms: [✅/❌]
6. Use Case Oriented: [✅/❌]
7. Result-Oriented: [✅/❌]
8. Tool/Method Specific: [✅/❌/N/A]

**Quality:**
✅ [Strengths]
❌ [Weaknesses]

---

### 4. Documentation: [X/10]

**File Linkage:**
- Total files: [N]
- Linked files: [N]
- Orphan files: [N]
- Broken links: [N]

**Documentation Quality:**
✅ [Good documentation]
❌ [Missing or poor documentation]

---

### 5. Integration: [X/10]

**CORE Duplication:**
- [✅ No duplication / ❌ Duplication found]

**Agent Compatibility:**
- [✅ Compatible / ❌ Issues found]

---

### 6. Quality: [X/10]

**Progressive Disclosure:** [✅/❌]
**Naming Conventions:** [✅/❌]
**Template Compliance:** [✅/❌]
**Examples Present:** [✅/❌]

---

## Issues Identified

### Critical Issues (Must Fix)
1. [Issue 1 with specific details]
2. [Issue 2 with specific details]

### Major Issues (Should Fix)
1. [Issue 1 with specific details]
2. [Issue 2 with specific details]

### Minor Issues (Nice to Fix)
1. [Issue 1 with specific details]
2. [Issue 2 with specific details]

---

## Recommendations

### Immediate Actions
1. [Action 1 to fix critical issues]
2. [Action 2 to fix critical issues]

### Short-term Improvements
1. [Action 1 for major issues]
2. [Action 2 for major issues]

### Long-term Enhancements
1. [Suggestion 1 for improvement]
2. [Suggestion 2 for improvement]

---

## Compliance Status

**This skill is:**
- [✅ COMPLIANT] - Meets all mandatory requirements
- [⚠️ PARTIALLY COMPLIANT] - Meets most requirements, minor issues
- [❌ NON-COMPLIANT] - Fails mandatory requirements

**Next Steps:**
[What needs to be done to achieve/maintain compliance]

---

## Canonical Reference

All validation based on:
`${PAI_DIR}/skills/CORE/skill-structure.md`

**Last Updated:** [YYYY-MM-DD]

Output: Complete validation report


Step 10: Present Results to User

Summarize findings:

  • Overall score and status
  • Critical issues requiring immediate attention
  • Recommendations for improvement
  • Next steps

If COMPLIANT: "✅ Skill [skill-name] is COMPLIANT with canonical architecture. Score: [X/70] ([percentage]%)"

If NON-COMPLIANT: "❌ Skill [skill-name] has compliance issues. Score: [X/70] ([percentage]%). [N] critical issues found. Recommend using canonicalize-skill workflow to fix."

Output: User informed of validation results


Success Criteria

Validation is complete when:

  • ✅ All 6 validation categories checked
  • ✅ Scores calculated for each category
  • ✅ Overall score computed
  • ✅ Issues identified and categorized
  • ✅ Recommendations provided
  • ✅ Compliance status determined
  • ✅ Report generated
  • ✅ User informed

Skill is COMPLIANT when:

  • Score ≥ 60/70 (≥85%)
  • No critical issues
  • Workflow Routing section present and FIRST
  • All workflows routed

Common Validation Failures

Failure 1: Missing Workflow Routing Section

Symptom: No "Workflow Routing" section in SKILL.md Impact: CRITICAL - Workflows never invoked Fix: Add Workflow Routing section FIRST, route all workflows

Failure 2: Orphan Workflows

Symptom: Workflow files exist but not routed in SKILL.md Impact: CRITICAL - Workflows inaccessible Fix: Add routing for each orphan workflow

Failure 3: Workflow Routing Not First

Symptom: Routing section buried in middle/end of SKILL.md Impact: MAJOR - Workflows may be missed Fix: Move Workflow Routing section to FIRST position

Failure 4: Incomplete Activation Triggers

Symptom: Only 2-3 categories covered, missing action verbs or modifiers Impact: MAJOR - Skill won't activate with common phrasings Fix: Expand to cover all 8 categories

Failure 5: Vague Examples

Symptom: Examples like "user asks for [skill]" instead of actual phrases Impact: MAJOR - Pattern matching fails Fix: Use real user phrasings, not templates

Failure 6: Unlinked Files

Symptom: Files in skill directory not referenced in SKILL.md Impact: MINOR - Files undiscoverable Fix: Link all files from SKILL.md main body

Failure 7: Wrong Archetype

Symptom: Complex structure with 5 workflows, or Minimal with 20 Impact: MINOR - Over/under-engineered Fix: Refactor to appropriate archetype


Related Workflows

  • create-skill.md - Create new compliant skill
  • canonicalize-skill.md - Fix non-compliant existing skill
  • update-skill.md - Update skill while maintaining compliance

Notes

Validation Philosophy:

  • Objective criteria (not subjective opinion)
  • Based on canonical architecture
  • Actionable recommendations
  • Clear pass/fail for critical items

Use Cases:

  • Pre-deployment quality gate
  • Post-creation verification
  • Periodic skill audits
  • Skill maintenance

One Source of Truth: ${PAI_DIR}/skills/CORE/skill-structure.md


Last Updated: 2025-11-17

Didn't find tool you were looking for?

Be as detailed as possible for better results