Axiomata Skill Evaluator v3.0#

Universal, impersonal skill quality evaluator for OpenClaw agents.

InfoValueVersion3.0.0TypeSelf-contained evaluation systemEvaluationDual (Axioma 5-Dim + ISO 25010)Target70+ (Axioma), 90%+ (ISO 25010)

---

1. Purpose and Scope#

Objective#

Provide complete skill quality evaluation using dual evaluation system:

1. Axioma 5-Dimension Framework (100 points)
2. ISO 25010 Structural Framework (13 automated checks)

Design Principles#

text

SELF-CONTAINED + UNIVERSAL + IMPERSONAL

PrincipleDescriptionSelf-containedAll tools bundled in skill directoryUniversalWorks for any OpenClaw agentImpersonalNo agent-specific references

When to Use#

TriggerActionEvaluate skillRun dual evaluationPre-publish checkRun full evaluation pipelineImprove skillAnalyze report and fix issuesSkill auditRun complete auditCheck qualityRun automated checks

---

2. Dual Evaluation System#

text

╔═══════════════════════════════════════════════════════════╗
║              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#

text

[INPUT] Skill to evaluate
          ↓
 Phase 1: Axioma 5-Dim Evaluation
          ↓
 Phase 2: ISO 25010 Structural Checks
          ↓
 [OUTPUT] Quality report + recommendations

---

3. Bundled Tools#

FileSystemPurposeevaluator.pyAxioma 5-Dim5-dimension evaluation, bundledeval-skill.pyISO 25010Automated structural checks, bundled

Tool Paths#

bash

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#

DimensionMaxTargetWhat it checksStructure2014+Header, sections, formatting, metadataClarity2014+Description, commands, examples, constraintsCompleteness2014+Tools, prerequisites, errors, edge casesConsistency2014+Naming, format, structure, styleFunctionality2014+Valid commands, documented outputs

4.2 Score Thresholds#

ScoreStatusAction90-100EXCELLENTReady for production70-89GOODMinor improvements possible50-69NEEDS_WORKMajor improvements needed<50POORSignificant rewrite required

---

5. ISO 25010 Structural Checks#

5.1 Automated Checks (13 total)#

CategoryChecksWhat it verifiesStructure6Frontmatter, header, sections, formattingTrigger2Description length, trigger contextsDocumentation3Body length, references, linked docsScripts2Python parse, no external deps

5.2 Pass Threshold#

ResultMeaning13/13Perfect structural score11-12/13Acceptable, minor warnings<11/13Needs structural improvement

---

6. Command Reference#

6.1 Axioma 5-Dim Evaluation#

bash

# 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#

bash

# 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#

bash

# 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#

text

=== 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#

json

{
  "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#

bash

$ 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#

bash

$ 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#

bash

$ 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:

bash

# 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%:

text

[{DIMENSION}] Low score ({percentage}%)
  -> {specific issue}
  -> {specific issue}

Example:

text

[CONSISTENCY] Low score (50%)
  -> Cluster alignment partial: 2
  -> Style: 2/5

---

11. Constraints#

ConstraintDescriptionPriority70% minimumSkills must score 70+ on Axioma 5-DimHIGH90%+ structuralTarget 90%+ on ISO 25010 checksHIGHSelf-containedAll tools bundled, no external depsHIGHImpersonalNo agent-specific referencesHIGH

Quality Thresholds#

SystemMinimumTargetAxioma 5-Dim70/10090/100ISO 2501011/1313/13Structure14/2018/20Clarity14/2018/20Completeness14/2018/20Consistency14/2018/20Functionality14/2018/20

---

12. Error Handling#

ErrorCauseSolutionSKILL.md not foundMissing fileCreate SKILL.md firstParse errorCorrupt fileCheck file encodingModule not foundMissing dependenciesUse bundled tools only

---

13. Workflow Integration#

text

╔═══════════════════════════════════════════════════════════╗
║         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