Battle Test

Battle Test Skill

[!IMPORTANT] Deep audit of a skills directory against the Skill Creator standard. Produces a scored report and phased remediation plan.

Optional args: slug=, ticket=<id/url>, mode=interactive|autonomous|channel, channel=, auto_continue=true|false, profile=business|hybrid|technical.

Instructions

When the user asks to perform this workflow, execute the following steps:

⚔️ Battle Test Orchestrator

Goal: Evaluate every SKILL.md in the target directory against common-skill-creator. Deliver a quantified health report and prioritized remediation plan.

---

Step 1 — Target Discovery & Tech Stack

Identify the tech stack and all skill files.

# Count total skills per category
find . -name "SKILL.md" | sed 's|/[^/]*/SKILL.md||' | sort | uniq -c

---

Step 2 — Frontmatter Audit (Breadth Scan)

Run scans to detect format and structure violations.

1. Check for missing mandatory sections: grep -rL "triggers:\|priority:\|Anti-Patterns" <SKILLS>/
2. Check for broad glob triggers: grep -r "src/\*\*" <SKILLS>/
3. Check for length limits: find . -name "SKILL.md" -exec awk 'END{if(NR>100) print FILENAME": "NR" lines"}' {} \;

---

Step 3 — Deep Audit & Evidence Classification

Evaluate skills against <SKILLS>/common/common-skill-creator/references/rubric.md:

1. Structural Checks: Line count (≤100), frontmatter schema, surgical triggers.
2. Textual Utility: Actionability, procedural clarity; verify NO answer-anchor padding.
3. Executable Outcomes: Runnable verification commands, deterministic test/verifier pass.
4. Rule Retirement & Ablation: Identify retirement candidates via repeated ablation; preserve safety controls until deterministic runtime replacements exist.

---

Step 4 — Scored Report

Scoring Algorithm: Start at 100 points for each category. Apply deductions for findings (🔴-15 / 🟠-8 / 🟡-3 / 🔵-1). Deduct for answer-anchor padding or bloated narrative.

📊 Report Format

Output the report using the Battle Test Report and Phased Plan templates in: <SKILLS>/common/common-skill-creator/references/rubric.md when synced.

Step 5 — Interactive Follow-up

1. "Generate a task.md for Phase 1 remediation?"
2. "Fix the worst offender in [category] now?"
3. "Deep-dive audit on a specific category (e.g., security)?"