skill-forge
Skill Forge
Generate Skills from requirements AND generate task_suite.yaml for evaluation.
The primary value of this skill is task_suite generation -- turning a SKILL.md into a structured test harness that improvement-evaluator can run. Secondary value is generating SKILL.md from a structured skill_spec.yaml.
Key differentiator: Skill Forge does not merely scaffold a skeleton; it performs static analysis of the SKILL.md to extract testable claims (from five distinct sources) and assigns the appropriate judge type per task, producing a suite that is immediately runnable by improvement-evaluator.
When to Use
- Add tests to an existing Skill -- You have a SKILL.md but no
task_suite.yaml. Run--from-skillto analyze the SKILL.md and generate a test suite automatically. The generator extracts scenarios from five sections of the document and assigns the right judge per task. - Create a new Skill from scratch -- You have a
skill_spec.yamldescribing what the skill should do. Run--from-specto generate both a complete SKILL.md (with frontmatter, sections, examples) and a matchingtask_suite.yaml. - Full lifecycle -- Generate a skill, test it with
--evaluate, and optionally improve it with--auto-improvein one pass. Combines skill-forge,
More from lanyasheng/auto-improvement-orchestrator-skill
skill-distill
|
1improvement-gate
当执行完变更需要验证是否应保留、候选被标记 pending 需要人工审批、或想查看待审队列时使用。6 层机械门禁: Schema→Compile→Lint→Regression→Review→HumanReview，其中 Schema/Compile/Regression/Review 为阻塞层（失败即拒绝），Lint 和 HumanReview 为建议层（失败不阻塞但记录警告）。不用于打分（用 improvement-discriminator）或执行变更（用 improvement-executor）。
1prompt-hardening
硬化 agent prompt、system prompt、SOUL.md、AGENTS.md、cron prompt 使 LLM 可靠遵循指令。触发词：agent 不听话、忽略规则、绕过约束、prompt 优化、指令合规、规则强化、prompt 硬化、LLM 不遵守、模型违规、creative circumvention。Use when agent ignores rules, disobeys instructions, bypasses tool constraints, needs prompt optimization, instruction compliance improvement, or rule hardening. 不适用于代码生成、代码审查、测试编写等执行型任务。参见 improvement-orchestrator (用于 skill 质量改进)、code-review-enhanced (用于代码审查)。
1benchmark-store
当需要初始化基准数据库、对比 skill 评分与历史基线、查看 Pareto front 是否有维度回退、或查阅质量分级标准时使用。不用于给候选打分（用 improvement-discriminator）或自动改进（用 improvement-learner）。
1improvement-evaluator
当需要验证 Skill 改进是否真正提升了 AI 执行效果时使用。通过预定义任务集（YAML）运行 AI 任务，判定 pass/fail，输出 execution_pass_rate。不用于文档结构评分（用 improvement-learner）或候选打分（用 improvement-discriminator）。
1improvement-learner
当需要检查 skill 质量评分、自动优化 SKILL.md 结构、追踪评估分数变化趋势、或「评分低了想知道哪里扣分」时使用。6维结构评估 + HOT/WARM/COLD 三层记忆 + Pareto front。不用于候选语义打分（用 improvement-discriminator）或全流程编排（用 improvement-orchestrator）。
1