codemap
Generate, update, or drift-check agent-facing CodeMaps as progressive code terrain indexes for projects, features, capabilities, functions, modules, or bug…
将 SDD-RIPER 方法论落地为严格可执行流程的重型 Harness 技能。用于辅助用户澄清最终目标、生成 codemap/context、拆分最小混沌单元、维护完整 spec、执行 RIPER 阶段门禁、阻塞高风险动作和沉淀可 new chat 恢复的本地任务轨迹。
$ npx -y skills add huisezhiyin/sdd-riper --skill sdd-riper-one --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/sdd-riper-oneContext preview
The summary Claude sees to decide when to auto-load this skill.
将 SDD-RIPER 方法论落地为严格可执行流程的重型 Harness 技能。用于辅助用户澄清最终目标、生成 codemap/context、拆分最小混沌单元、维护完整 spec、执行 RIPER 阶段门禁、阻塞高风险动作和沉淀可 new chat 恢复的本地任务轨迹。
name: sdd-riper-one description: 将 SDD-RIPER 方法论落地为严格可执行流程的重型 Harness 技能。用于辅助用户澄清最终目标、生成 codemap/context、拆分最小混沌单元、维护完整 spec、执行 RIPER 阶段门禁、阻塞高风险动作和沉淀可 new chat 恢复的本地任务轨迹。
在 checkpoint、执行前、阶段收尾、偏差暴露或 new chat 恢复时,先让模型用当前上下文自总结:
热摘要只用于当前轮聚焦,不替代 spec;与 spec 冲突时,永远以 spec 为准。它是比完整 spec 更轻的 recap checkpoint,用来防止当前 loop 上下文腐烂;除 checkpoint/恢复/高风险节点外,不机械重复这组字段。
Code is cheap. Context, control, and evidence are not. Let the model move the work forward. Let the human own goals, boundaries, permissions, checkpoints, evidence, and acceptance. Strong coding agents no longer stop at suggestions.
Generate, update, or drift-check agent-facing CodeMaps as progressive code terrain indexes for projects, features, capabilities, functions, modules, or bug…
Prepare safe, compact new-chat handoffs when a user wants a fresh chat, resume pack, context reset, recovery, or agent handoff. Before creating a new…
面向 GPT-5.4 等强模型和熟练用户的轻量 AI Agent Harness / checkpoint-driven coding skill。默认用户已经把任务切到基本可执行的最小混沌单元;模型自行分解、探索与推进,人类通过最终目标、最小 spec、复述、checkpoint、证据验证与回写来低干扰控盘。