ynh-sensors
Choose, declare and calibrate the sensors a project needs, then gate on them with ynh check. Use when a harness has no sensors, when ynh check reports nothing useful, or when setting up an agent loop that needs something to converge against.
原文 · 译文尚未补齐适用对象:对象待分类。自动收录 · 待核查。
原作者说明节选
从固定版本的 SKILL.md 提取,保留原文语言。以下内容属于作者说明,本站尚未验证安装、兼容或运行结果。
查看英文原文
原文涵盖的主题
- Set up sensors for a project
- Before you start
- The vocabulary is a closed set
- Step 1 — What is already there?
- Step 2 — Propose a set across all three categories
- Step 3 — Place each on the tolerance ladder
- Step 4 — Adopt into a repo that is already failing
- Step 5 — For a files sensor, declare observes
You are helping someone decide what their project should observe, declare it in their harness, and make ynh check a gate they trust. Work through this one step at a time. Ask, wait for the answer, then move on — do not dump the whole plan at once. references/starter-sets.md — proposed sensor sets per stack (Go, Node/TS, Python, Rust, JVM, polyglot monorepo) references/calibration.md — reference fixtures, ratchets, freshness and versioncommand, with the reasoning for each Both ship with this skill. Run ynh sensors --help and ynh check --help for the live CLI — they describe the installed version and run nothing. Get these wrong and ynd validate rejects the manifest, so establish them first rather than discovering them: | Field | Values | Meaning | |---|---|---| | category | maintainability · architecture · behaviour | What kind of health this observes | | tolerance | blocking (default) · advisory · report | What a failure does | | source | exactly one of files · command · focus | How the observation is made | | role | regular · convergence-verifier · stuck-recovery | How a loop driver finds it | | ratchet | fingerprint (default) · count | How the baseline forgives | is an installed
这里只展示节选,请以完整原文和当前要求为准。
- 来源仓库
- eyelock/ynh
- 源文件
- skills/ynh-sensors/SKILL.md
- 固定版本
- 3a2944a97f6341a43490dac4e5b1fd37e8c3b6e6
准备条件
- 本条仅收录原作者的技能元数据;请阅读完整说明,确认依赖、许可与使用条件。
第一件可以尝试的小事
先阅读原作者说明,再决定是否在自己的 AI 工具中使用。
对应 AI 工具说明
作者有说明不等于本站实机验证。没有证据时,不自动标为兼容、安全或免费。
来源与内容状态
本条为自动发现的元数据,尚未完成人工文档核查;适用性与运行结果待核查。
原作者:eyelock/ynh · 原作者仓库元数据。文档核查:未核查。最近来源检查:2026-11-10T23:01:52.120Z。
分类由名称、说明与路径规则生成,分类不代表已核实适配。
许可与权利归属
MIT。第三方内容及版权归原作者,使用条件以原文为准。
许可原文 ↗本站没有因同步、复制或用户自报成功而将资源标为运行验证通过。
限制与注意事项
- 元数据收录不代表本站已经安装、适配、实测或进行安全审查。