Blind Scorer — Development agent for Claude Code
Blind quality scorer for skill-improver runs.
How to install Blind Scorer
Installs to ~/.claude/agents/air-gapped-skills-blind-scorer.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/air-gapped/skills/HEAD/.claude/agents/blind-scorer.md -o ~/.claude/agents/air-gapped-skills-blind-scorer.md Restart Claude Code, or start a new session, for it to be picked up.
What Blind Scorer does
name: blind-scorer description: Blind quality scorer for skill-improver runs. Spawn explicitly with a rubric dir and target dir — not for general delegation. tools: Read, Grep, Glob, Bash model: sonnet
Score one Claude Code skill against the skill-improver quality rubric. This system prompt is the complete, canonical scoring instruction set; the spawn prompt supplies only two paths:
RUBRIC DIR:— skill-improver'sreferences/directory (rubric + design guide)TARGET DIR:— the
Alternatives in Development
- General Code Quality Debugger — Use this agent proactively when you need systematic code review, debugging assistance, refactoring guidance, o 248 ★
- Part 13: Memory Bridge (Give Coding Agents Your Brain) — When you spawn Codex or Claude Code to build something, they start blind 232 ★
- Thermo Nuclear Code Quality Review — Independent strict maintainability reviewer for a collected diff and changed-file contents 144 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Skill Comparator
Blind A/B comparator for skill-improver runs. Spawn explicitly with a rubric dir and two unlabelled skill dirs
Fit Scorer
Scores every company on the target list against the buy box into PRIORITY, WATCH or PASS. Builds a written rub
Patch Author
Patch author for one verified vulnerability finding in a patch run. Spawn with a repo path and a nonce-tagged
Rfe Scorer
Scores a single RFE issue against quality rubric. Restricted to Read and Write only to prevent prompt injectio
Gauntlet Critic
Independent critic for one piece of a Gauntlet Loop round. Inspects real artifacts, compares them blind agains
Lane Implementer
Implementation lane with the full gate battery baked in - private target dir, all-targets clippy, targeted tes
Related Skills
Claude Code Game Builder
Claude Code skill: multi-agent build → capture → blind critic → fix loops that push a game toward a reference-
05 Verification Panel
<!-- Injected: {{ANGLE}} {{ANGLE_QUESTION}} {{TARGETS_JSON}} {{PAPER_TXT_PATH}} {{STAGED_SOURCES_DIR}} {{QUOTE
QA Sync
Sync generated/healed tests from the plugin output dir into the target app repo.