Axiomata Skill Evaluator v3.0
Universal, impersonal skill quality evaluator for OpenClaw agents.
| Info |
Value |
| Version |
3.0.0 |
| Type |
Self-contained evaluation system |
| Evaluation |
Dual (Axioma 5-Dim + ISO 25010) |
| Target |
70+ (Axioma), 90%+ (ISO 25010) |
1. Purpose and Scope
Objective
Provide complete skill quality evaluation using dual evaluation system:
- Axioma 5-Dimension Framework (100 points)
- ISO 25010 Structural Framework (13 automated checks)
Design Principles
SELF-CONTAINED + UNIVERSAL + IMPERSONAL
| Principle |
Description |
| Self-contained |
All tools bundled in skill directory |
| Universal |
Works for any OpenClaw agent |
| Impersonal |
No agent-specific references |
When to Use
| Trigger |
Action |
| Evaluate skill |
Run dual evaluation |
| Pre-publish check |
Run full evaluation pipeline |
| Improve skill |
Analyze report and fix issues |
| Skill audit |
Run complete audit |
| Check quality |
Run automated checks |
2. Dual Evaluation System
╔═══════════════════════════════════════════════════════════╗
║ DUAL EVALUATION ARCHITECTURE ║
╠═══════════════════════════════════════════════════════════╣
║ ║
║ ┌─────────────────────────────────────────────────┐ ║
║ │ 1. Axioma 5-Dim Evaluation (100 pts) │ ║
║ │ │ ║
║ │ Dimensions: │ ║
║ │ ├─ Structure → 20% │ ║
║ │ ├─ Clarity → 20% │ ║
║ │ ├─ Completeness → 20% │ ║
║ │ ├─ Consistency → 20% │ ║
║ │ └─ Functionality → 20% │ ║
║ │ │ ║
║ │ Target: 70+ score │ ║
║ └─────────────────────────────────────────────────┘ ║
║ ↓ ║
║ ┌─────────────────────────────────────────────────┐ ║
║ │ 2. ISO 25010 Structural (13 checks) │ ║
║ │ │ ║
║ │ Categories: 8 categories, 25 standards │ ║
║ │ Automated checks: 13 tests │ ║
║ │ │ ║
║ │ Target: 90%+ (12/13 checks passing) │ ║
║ └─────────────────────────────────────────────────┘ ║
║ ║
╚═══════════════════════════════════════════════════════════╝
Evaluation Flow
[INPUT] Skill to evaluate
↓
Phase 1: Axioma 5-Dim Evaluation
↓
Phase 2: ISO 25010 Structural Checks
↓
[OUTPUT] Quality report + recommendations
3. Bundled Tools
| File |
System |
Purpose |
evaluator.py |
Axioma 5-Dim |
5-dimension evaluation, bundled |
eval-skill.py |
ISO 25010 |
Automated structural checks, bundled |
Tool Paths
EVAL_PATH="<skill-directory>/scripts/"
SKILL_PATH="<skill-to-evaluate>"
# Axioma 5-Dim evaluation
python3 ${EVAL_PATH}evaluator.py <SKILL_PATH> --verbose
# ISO 25010 checks
python3 ${EVAL_PATH}eval-skill.py <SKILL_PATH> --verbose
4. Axioma 5-Dimension Evaluation
4.1 Dimension Breakdown
| Dimension |
Max |
Target |
What it checks |
| Structure |
20 |
14+ |
Header, sections, formatting, metadata |
| Clarity |
20 |
14+ |
Description, commands, examples, constraints |
| Completeness |
20 |
14+ |
Tools, prerequisites, errors, edge cases |
| Consistency |
20 |
14+ |
Naming, format, structure, style |
| Functionality |
20 |
14+ |
Valid commands, documented outputs |
4.2 Score Thresholds
| Score |
Status |
Action |
| 90-100 |
EXCELLENT |
Ready for production |
| 70-89 |
GOOD |
Minor improvements possible |
| 50-69 |
NEEDS_WORK |
Major improvements needed |
| <50 |
POOR |
Significant rewrite required |
5. ISO 25010 Structural Checks
5.1 Automated Checks (13 total)
| Category |
Checks |
What it verifies |
| Structure |
6 |
Frontmatter, header, sections, formatting |
| Trigger |
2 |
Description length, trigger contexts |
| Documentation |
3 |
Body length, references, linked docs |
| Scripts |
2 |
Python parse, no external deps |
5.2 Pass Threshold
| Result |
Meaning |
| 13/13 |
Perfect structural score |
| 11-12/13 |
Acceptable, minor warnings |
| <11/13 |
Needs structural improvement |
6. Command Reference
6.1 Axioma 5-Dim Evaluation
# Basic evaluation
python3 <skill-directory>/scripts/evaluator.py <skill-path>
# Verbose output
python3 <skill-directory>/scripts/evaluator.py <skill-path> --verbose
# With auto-improvement suggestions
python3 <skill-directory>/scripts/evaluator.py <skill-path> --verbose --improve
# JSON output
python3 <skill-directory>/scripts/evaluator.py <skill-path> --json
6.2 ISO 25010 Checks
# Basic ISO check
python3 <skill-directory>/scripts/eval-skill.py <skill-path>
# Verbose output
python3 <skill-directory>/scripts/eval-skill.py <skill-path> --verbose
# JSON output
python3 <skill-directory>/scripts/eval-skill.py <skill-path> --json
6.3 Evaluate All Skills
# Evaluate all skills in parent directory
python3 <skill-directory>/scripts/evaluator.py --all
# Verbose all
python3 <skill-directory>/scripts/evaluator.py --all --verbose
7. Output Formats
7.1 Console Output
=== EVALUATION RESULTS ===
STRUCTURE 20/20 ████████████████████ 100%
CLARITY 17/20 █████████████████░░░ 85%
COMPLETENESS 14/20 ████████████░░░░░░░░ 70%
CONSISTENCY 10/20 ██████████░░░░░░░░░ 50%
FUNCTIONALITY 12/20 ████████████░░░░░░░░ 60%
------------------------------------------------------------
STATUS: GOOD (score 73%)
7.2 JSON Output
{
"skill": "example-skill",
"scores": {
"structure": 20,
"clarity": 17,
"completeness": 14,
"consistency": 10,
"functionality": 12
},
"total": 73,
"max": 100,
"passed": true,
"recommendations": [
"[CONSISTENCY] Low score (50%)",
" -> Style: 2/5"
]
}
8. Examples
Example 1: Basic Evaluation
$ python3 evaluator.py /path/to/skill --verbose
Read /path/to/skill/SKILL.md (2048 chars)
Structure: 20/20
Clarity: 17/20
Completeness: 14/20
Consistency: 10/20
Functionality: 12/20
============================================================
STATUS: GOOD (score 73%)
Example 2: ISO 25010 Check
$ python3 eval-skill.py /path/to/skill --verbose
[STRUCTURE]
Pass: 5/6
[TRIGGER]
Pass: 2/2
[DOCUMENTATION]
Pass: 2/3
[SCRIPTS]
Pass: 2/2
[SECURITY]
Pass: 2/2
==================================================
Pass: 13 Warn: 0 Fail: 0
Structural score: 100% (13/13 checks passed)
Example 3: JSON Output
$ python3 evaluator.py /path/to/skill --json
{
"skill": "example-skill",
"scores": {
"structure": 20,
"clarity": 17,
"completeness": 14,
"consistency": 10,
"functionality": 12
},
"total": 73,
"max": 100,
"passed": true,
"recommendations": []
}
9. Self-Evaluation
The skill evaluates itself automatically:
# Self-evaluation (when no path given, evaluates own directory)
python3 evaluator.py
# Expected output for v3.0:
# Score: 85+/100
# Status: GOOD to EXCELLENT
10. Recommendations Format
When a dimension scores below 70%:
[{DIMENSION}] Low score ({percentage}%)
-> {specific issue}
-> {specific issue}
Example:
[CONSISTENCY] Low score (50%)
-> Cluster alignment partial: 2
-> Style: 2/5
11. Constraints
| Constraint |
Description |
Priority |
| 70% minimum |
Skills must score 70+ on Axioma 5-Dim |
HIGH |
| 90%+ structural |
Target 90%+ on ISO 25010 checks |
HIGH |
| Self-contained |
All tools bundled, no external deps |
HIGH |
| Impersonal |
No agent-specific references |
HIGH |
Quality Thresholds
| System |
Minimum |
Target |
| Axioma 5-Dim |
70/100 |
90/100 |
| ISO 25010 |
11/13 |
13/13 |
| Structure |
14/20 |
18/20 |
| Clarity |
14/20 |
18/20 |
| Completeness |
14/20 |
18/20 |
| Consistency |
14/20 |
18/20 |
| Functionality |
14/20 |
18/20 |
12. Error Handling
| Error |
Cause |
Solution |
| SKILL.md not found |
Missing file |
Create SKILL.md first |
| Parse error |
Corrupt file |
Check file encoding |
| Module not found |
Missing dependencies |
Use bundled tools only |
13. Workflow Integration
╔═══════════════════════════════════════════════════════════╗
║ SKILL PUBLISHING WORKFLOW ║
╠═══════════════════════════════════════════════════════════╣
║ ║
║ 1. Create Skill ║
║ ↓ ║
║ 2. Self-Evaluate ║
║ ↓ 70+ → Continue ║
║ 3. ISO 25010 Check ║
║ ↓ 90%+ → Continue ║
║ 4. Fix Issues if Needed ║
║ ↓ ║
║ 5. Publish to ClawHub ║
║ ║
║ RESULT: QUALITY SKILLS FOR PRODUCTION ║
║ ║
╚═══════════════════════════════════════════════════════════╝
In Altum Per Qualitatem.
🧪 AXIOMATA SKILL EVALUATOR v3.0 — UNIVERSAL QUALITY SYSTEM
