Automates compliance measurement for coding agents by generating test scenarios from skill/rule files, running agents, classifying tool calls, and reporting compliance rates with timelines.
How this skill is triggered — by the user, by Claude, or both
Slash command
/everything-claude-code:skill-complyThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
衡量编码智能体是否真正遵循技能、规则或智能体定义,通过:
fixtures/compliant_trace.jsonlfixtures/noncompliant_trace.jsonlfixtures/tdd_spec.yamlprompts/classifier.mdprompts/scenario_generator.mdprompts/spec_generator.mdpyproject.tomlscripts/__init__.pyscripts/classifier.pyscripts/grader.pyscripts/parser.pyscripts/report.pyscripts/run.pyscripts/runner.pyscripts/scenario_generator.pyscripts/spec_generator.pyscripts/utils.pytests/test_grader.pytests/test_parser.pytests/test_runner.py衡量编码智能体是否真正遵循技能、规则或智能体定义,通过:
claude -p 并通过 stream-json 捕获工具调用追踪skills/*/SKILL.md):工作流技能如 search-first、TDD 指南rules/common/*.md):强制规则如 testing.md、security.md、git-workflow.mdagents/*.md):智能体是否在预期时被调用(内部工作流验证尚未支持)/skill-comply <path># 完整运行
uv run python -m scripts.run ~/.claude/rules/common/testing.md
# 试运行(无成本,仅规格 + 场景)
uv run python -m scripts.run --dry-run ~/.claude/skills/search-first/SKILL.md
# 自定义模型
uv run python -m scripts.run --gen-model haiku --model sonnet <path>
衡量即使提示没有明确支持某项技能/规则时,该技能/规则是否仍被遵循。
报告是自包含的,包括:
对于熟悉钩子的用户,报告还包括对合规率低的步骤的钩子推广建议。这是信息性的 — 主要价值在于合规可见性本身。
npx claudepluginhub aaione/everything-claude-code-zhMeasures whether skills, rules, and agent definitions are actually followed by auto-generating test scenarios at 3 strictness levels and reporting compliance rates with full tool call timelines.
Measures coding agent compliance with skills/rules/agents by generating specs/scenarios at 3 strictness levels, running agents, classifying tool calls, and reporting timelines with scores.
Auto-generates test scenarios from CLAUDE.md rules, runs them with LLM-as-judge scoring, reports compliance score and per-rule breakdown. Optionally improves failing rules via mutation loop.