bc7score — How Your Skill Gets Scored
Every skill submitted to the Hordago marketplace is automatically scored by the bc7 scoring engine. Your score determines:
- Whether your skill is accepted (minimum: 60 / 100)
- How prominently it appears in search results
- Whether it earns a Featured or Verified badge
Score Breakdown
| Dimension | Weight | What it measures |
|---|---|---|
| Trigger clarity | 25% | How precisely the trigger describes when to use the skill |
| Instruction quality | 25% | Are the body instructions specific, actionable, and complete? |
| Frontmatter completeness | 20% | All required + recommended fields present |
| Tag relevance | 10% | Tags match the skill’s actual behavior |
| Description conciseness | 10% | Description fits in one clear sentence |
| Example coverage | 10% | Does the body include at least one example? |
Scoring Tiers
| Score | Tier | Result |
|---|---|---|
| 90–100 | Platinum | Featured in homepage carousel |
| 80–89 | Gold | Boosted in search rankings |
| 70–79 | Silver | Standard listing |
| 60–69 | Bronze | Listed, no boost |
| < 60 | Rejected | Returned with feedback |
How to Improve Your Score
Trigger clarity (most impactful)
A vague trigger scores low. Be specific about the conditions.
# Low score (vague)
trigger: "Use when the user needs help."
# High score (specific)
trigger: "Use when the user asks to generate a research protocol, experimental design, or lab procedure document."Instruction quality
The body should tell the agent exactly what to do, in what order, with what output format.
# Bad — too vague
Help the user with their task.
# Good — specific and structured
1. Ask for the experiment type (knockout, base editing, CRISPRa/i)
2. Ask for the target gene and cell line
3. Generate a protocol using the appropriate template
4. Validate reagent availability before finalizingExample coverage
Include at least one concrete example of inputs and expected output. The scorer checks for example blocks or > blockquotes in the body.
Running the Scorer Locally
npx @hordago/skill-validator ./my-skill/SKILL.md --verboseThe --verbose flag shows per-dimension scores so you know exactly what to fix.