[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EXDR4HNXMM66VPME60Q63Q":3,"skill-same:01M7EXDR4HNXMM66VPME60Q63Q":47},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":30,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M7EXDR4HNXMM66VPME60Q63Q","ask-the-human",null,"Use when you hit an open question, a decision that is not yours to make, a gap or ambiguity in a spec, or a draft that needs a human's confirmation. Asks through the host's structured-question tool with concrete options and a recommendation, instead of guessing or burying the question in prose.","IntraverseGame\u002Fse-client-kit","IntraverseGame","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F106401126?v=4","01M7EXDHS7CGXX7G1CM6ZR69AG","Intraverse","01M7EXDHSXGC8ZJRQY5E0XCH6H",2,"MIT","https:\u002F\u002Fgithub.com\u002FIntraverseGame\u002Fse-client-kit\u002Fblob\u002Fmain\u002F.ai\u002Fskills\u002Fask-the-human\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: ask-the-human\ndescription: Use when you hit an open question, a decision that is not yours to make, a gap or ambiguity in a spec, or a draft that needs a human's confirmation. Asks through the host's structured-question tool with concrete options and a recommendation, instead of guessing or burying the question in prose.\n---\n\n# ask-the-human\n\nThe one way every role agent raises a question to the human. It is\ntool-agnostic: it names the capability (a structured question with selectable\noptions) and says what to do on a host that lacks it.\n\n## Prerequisites\n\n- You know what is blocking you and have already looked it up, in the\n  precedence order of [AGENTS.md](..\u002F..\u002F..\u002FAGENTS.md#precedence-and-secrets):\n  code\u002Fconfig, specs, decisions, runbook, memories. Ask only what they cannot\n  answer.\n\n## Workflow\n\n1. **Decide whether to ask.**\n   - **Ask** when: the spec is missing or ambiguous; code and spec disagree and\n     the intended behavior is unclear; two valid designs have different\n     consequences; scope or acceptance criteria are unclear; an RFC or ADR\n     still has an undecided item; a draft needs the human's confirmation; an\n     action is hard to reverse or outward-facing; something belongs to someone\n     else (changes in another worktree).\n   - **Do not ask** when: the repository answers it; a convention gives an\n     obvious default (take it and say so in your report); it is a permission\n     ping (\"Should I proceed?\", \"Is the plan ready?\"); or an approval gate\n     already covers it.\n2. **Collect first.** Gather every open question before asking, and ask them\n   in one batch. A question that depends on another's answer waits for it. If\n   there are more than the tool takes per call, ask the ones that block you\n   first and the rest in a second call; do not drop any.\n3. **Shape each question.**\n   - One decision, self-contained, ending in `?`, with a short header naming\n     the topic.\n   - A few options, mutually exclusive unless the question is multi-select (say\n     so in the question). Each option has a short label and a description of\n     what happens if it is chosen.\n   - **Recommend only with a basis.** When a spec, ADR, accepted convention, or\n     the surrounding code favors one option, put it first, mark it as\n     recommended, and cite the source in its description\n     (`specs\u002F\u003Cfile>.md#\u003Csection>`, `ADR-NNN`). With no basis, recommend\n     nothing; never invent one.\n   - When options differ in something concrete (a snippet, a config, a layout),\n     show it in the option's preview if the tool supports previews.\n4. **Ask through the tool.**\n   - **Claude Code:** `AskUserQuestion`. Its limits: up to 4 related questions\n     per call, 2 to 4 options each, a header of 12 characters or fewer, a\n     label of 1 to 5 words, `(Recommended)` at the end of the recommended\n     label, and a free-text \"Other\" it adds itself, so do not add one.\n   - **Another host:** the equivalent structured-question tool in your tool\n     list, within its limits.\n   - **No such tool, or you are a subagent that cannot prompt the human:** do\n     not continue past the blocker on a guess. In chat, print the same\n     questions as a numbered list with lettered options and the recommended one\n     marked, and stop. A subagent returns them under **Open questions** in its\n     final report so the parent agent asks.\n5. **Act on the answer.**\n   - Apply it and say what you decided in your report: question, answer, where\n     it is recorded.\n   - Record durable answers where they belong: behavior in the spec (or an\n     RFC\u002FADR, through the [documenter](..\u002F..\u002Fagents\u002Fdocumenter.md)); in-flight\n     context in a [working memory](..\u002F..\u002Fmemories\u002Fworking\u002FREADME.md).\n   - An item the human defers, or leaves unanswered and that does not block\n     you, becomes a `\u003C!-- TODO -->` marker; never fill it in yourself. An item that\n     does block you stops you (step 4).\n   - A free-text answer is the human's decision, not a hint: follow it. If it\n     contradicts an accepted spec or ADR, say so in your report and update the\n     spec in the same change (ADR-002; the implementer hands that to the\n     [documenter](..\u002F..\u002Fagents\u002Fdocumenter.md)); if an accepted ADR is affected,\n     open an RFC or supersede it with a new ADR rather than leaving it stale.\n     Explicit developer instructions outrank specs and decisions (AGENTS.md\n     precedence).\n   - \"You decide\" means take the recommended option, or your best-grounded one,\n     and state it.\n\n## Constraints\n\n- **An answer is not an approval gate.** It never stands in for the human's\n  review of a diff or the reviewer agent's approval (ADR-003). Do not use it to\n  ask whether to stage, commit, or push: that waits for the human's own\n  request.\n- **Outward-facing text is shown in full first.** For anything posted outside\n  the repository (pull request comments, ADR-003), print the exact text in\n  chat, then ask with options such as _Post as written_, _Change it_, _Do not\n  post_. This repository is public, so what is posted is world-readable. A confirmation covers that text only; a changed draft is asked again.\n- Approval in one context does not extend to another: ask again for each\n  hard-to-reverse or outward-facing action.\n- One round of questions (it may take more than one call), then proceed. Do\n  not ask the same question twice or keep a conversation going through the\n  tool.\n- Questions and options carry no secrets, credentials, or internal hostnames.\n  Treat chat output as possibly pasted into a public issue.\n\n## Example\n\nIllustrative; the placeholders stand for real paths and ADR numbers.\n\n```text\nheader:   Retry policy\nquestion: Which retry policy should the \u003Cjob> use when \u003Cdependency> times out?\noptions:\n  1. Fixed delay (Recommended)\n       specs\u002F\u003Cfile>.md#\u003Csection> already uses it for \u003Csimilar job>.\n  2. Exponential backoff\n       Better for a long outage; adds a config variable and a spec change.\n  3. No retry\n       Fails the round at once; the \u003Ccaller> must handle it.\n```\n\n## Expected output\n\nThe answers applied, and a short list in your report of each decision: the\nquestion, the answer, and where it is recorded. Items still open are listed as\n`TODO`s.\n\n## References\n\n- [AGENTS.md](..\u002F..\u002F..\u002FAGENTS.md#asking-the-human),\n  [CONTRIBUTING.md#asking-the-human](..\u002F..\u002F..\u002FCONTRIBUTING.md#asking-the-human)\n- [ADR-003](..\u002F..\u002F..\u002Fdecisions\u002FADR-003-agent-workflow-and-approvals.md)\n","# ask-the-human\n\nThe one way every role agent raises a question to the human. It is\ntool-agnostic: it names the capability (a structured question with selectable\noptions) and says what to do on a host that lacks it.\n\n## Prerequisites\n\n- You know what is blocking you and have already looked it up, in the\n  precedence order of [AGENTS.md](..\u002F..\u002F..\u002FAGENTS.md#precedence-and-secrets):\n  code\u002Fconfig, specs, decisions, runbook, memories. Ask only what they cannot\n  answer.\n\n## Workflow\n\n1. **Decide whether to ask.**\n   - **Ask** when: the spec is missing or ambiguous; code and spec disagree and\n     the intended behavior is unclear; two valid designs have different\n     consequences; scope or acceptance criteria are unclear; an RFC or ADR\n     still has an undecided item; a draft needs the human's confirmation; an\n     action is hard to reverse or outward-facing; something belongs to someone\n     else (changes in another worktree).\n   - **Do not ask** when: the repository answers it; a convention gives an\n     obvious default (take it and say so in your report); it is a permission\n     ping (\"Should I proceed?\", \"Is the plan ready?\"); or an approval gate\n     already covers it.\n2. **Collect first.** Gather every open question before asking, and ask them\n   in one batch. A question that depends on another's answer waits for it. If\n   there are more than the tool takes per call, ask the ones that block you\n   first and the rest in a second call; do not drop any.\n3. **Shape each question.**\n   - One decision, self-contained, ending in `?`, with a short header naming\n     the topic.\n   - A few options, mutually exclusive unless the question is multi-select (say\n     so in the question). Each option has a short label and a description of\n     what happens if it is chosen.\n   - **Recommend only with a basis.** When a spec, ADR, accepted convention, or\n     the surrounding code favors one option, put it first, mark it as\n     recommended, and cite the source in its description\n     (`specs\u002F\u003Cfile>.md#\u003Csection>`, `ADR-NNN`). With no basis, recommend\n     nothing; never invent one.\n   - When options differ in something concrete (a snippet, a config, a layout),\n     show it in the option's preview if the tool supports previews.\n4. **Ask through the tool.**\n   - **Claude Code:** `AskUserQuestion`. Its limits: up to 4 related questions\n     per call, 2 to 4 options each, a header of 12 characters or fewer, a\n     label of 1 to 5 words, `(Recommended)` at the end of the recommended\n     label, and a free-text \"Other\" it adds itself, so do not add one.\n   - **Another host:** the equivalent structured-question tool in your tool\n     list, within its limits.\n   - **No such tool, or you are a subagent that cannot prompt the human:** do\n     not continue past the blocker on a guess. In chat, print the same\n     questions as a numbered list with lettered options and the recommended one\n     marked, and stop. A subagent returns them under **Open questions** in its\n     final report so the parent agent asks.\n5. **Act on the answer.**\n   - Apply it and say what you decided in your report: question, answer, where\n     it is recorded.\n   - Record durable answers where they belong: behavior in the spec (or an\n     RFC\u002FADR, through the [documenter](..\u002F..\u002Fagents\u002Fdocumenter.md)); in-flight\n     context in a [working memory](..\u002F..\u002Fmemories\u002Fworking\u002FREADME.md).\n   - An item the human defers, or leaves unanswered and that does not block\n     you, becomes a `\u003C!-- TODO -->` marker; never fill it in yourself. An item that\n     does block you stops you (step 4).\n   - A free-text answer is the human's decision, not a hint: follow it. If it\n     contradicts an accepted spec or ADR, say so in your report and update the\n     spec in the same change (ADR-002; the implementer hands that to the\n     [documenter](..\u002F..\u002Fagents\u002Fdocumenter.md)); if an accepted ADR is affected,\n     open an RFC or supersede it with a new ADR rather than leaving it stale.\n     Explicit developer instructions outrank specs and decisions (AGENTS.md\n     precedence).\n   - \"You decide\" means take the recommended option, or your best-grounded one,\n     and state it.\n\n## Constraints\n\n- **An answer is not an approval gate.** It never stands in for the human's\n  review of a diff or the reviewer agent's approval (ADR-003). Do not use it to\n  ask whether to stage, commit, or push: that waits for the human's own\n  request.\n- **Outward-facing text is shown in full first.** For anything posted outside\n  the repository (pull request comments, ADR-003), print the exact text in\n  chat, then ask with options such as _Post as written_, _Change it_, _Do not\n  post_. This repository is public, so what is posted is world-readable. A confirmation covers that text only; a changed draft is asked again.\n- Approval in one context does not extend to another: ask again for each\n  hard-to-reverse or outward-facing action.\n- One round of questions (it may take more than one call), then proceed. Do\n  not ask the same question twice or keep a conversation going through the\n  tool.\n- Questions and options carry no secrets, credentials, or internal hostnames.\n  Treat chat output as possibly pasted into a public issue.\n\n## Example\n\nIllustrative; the placeholders stand for real paths and ADR numbers.\n\n```text\nheader:   Retry policy\nquestion: Which retry policy should the \u003Cjob> use when \u003Cdependency> times out?\noptions:\n  1. Fixed delay (Recommended)\n       specs\u002F\u003Cfile>.md#\u003Csection> already uses it for \u003Csimilar job>.\n  2. Exponential backoff\n       Better for a long outage; adds a config variable and a spec change.\n  3. No retry\n       Fails the round at once; the \u003Ccaller> must handle it.\n```\n\n## Expected output\n\nThe answers applied, and a short list in your report of each decision: the\nquestion, the answer, and where it is recorded. Items still open are listed as\n`TODO`s.\n\n## References\n\n- [AGENTS.md](..\u002F..\u002F..\u002FAGENTS.md#asking-the-human),\n  [CONTRIBUTING.md#asking-the-human](..\u002F..\u002F..\u002FCONTRIBUTING.md#asking-the-human)\n- [ADR-003](..\u002F..\u002F..\u002Fdecisions\u002FADR-003-agent-workflow-and-approvals.md)\n",{"name":5,"description":7},"main","4b537f4ee2a8df7dba4156583d62922c0c87399c",".ai\u002Fskills\u002Fask-the-human",0,"https:\u002F\u002Fgithub.com\u002FIntraverseGame\u002Fse-client-kit","https:\u002F\u002Fgithub.com\u002FIntraverseGame","https:\u002F\u002Fplay.intraverse.io","Intraverse is a platform creating engagement through cross-community.","2026-11-14 02:21:01",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add IntraverseGame\u002Fse-client-kit --skill ask-the-human",{"tool":39,"command":40},"bunx","bunx skills add IntraverseGame\u002Fse-client-kit --skill ask-the-human",{"tool":42,"command":43},"pnpm","pnpm dlx skills add IntraverseGame\u002Fse-client-kit --skill ask-the-human","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FIntraverseGame\u002Fse-client-kit\u002Fblob\u002Fmain\u002F.ai\u002Fskills\u002Fask-the-human\u002FSKILL.md",true,[],[48,55,62,69,76,83],{"skillNo":49,"name":50,"nameZh":6,"description":51,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":52,"categories":53,"keywords":54,"publishedAt":19},"01M7EXECT0MKSVJ78TDP62R8CZ","\u003Cskill-name>","Use when \u003Csituation>. One or two sentences.","https:\u002F\u002Fgithub.com\u002FIntraverseGame\u002Fse-client-kit\u002Fblob\u002Fmain\u002F.ai\u002Ftemplates\u002Fskill.md",[],[],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":19},"01M7EXE9STPWRCHBKPMRXBA4JE","write-tests","Use when adding or changing behavior, signing code, contracts, or the UI. Chooses the right test level and applies the repository's test rules.","https:\u002F\u002Fgithub.com\u002FIntraverseGame\u002Fse-client-kit\u002Fblob\u002Fmain\u002F.ai\u002Fskills\u002Fwrite-tests\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":19},"01M7EXE5SGQEC80S64WMVGKRCZ","spec-drift-sync","Use before merging any non-trivial change. Reconciles code with specs, and code\u002Fspecs with memories, runbooks, and telemetry docs (Spec Drift Synchronization).","https:\u002F\u002Fgithub.com\u002FIntraverseGame\u002Fse-client-kit\u002Fblob\u002Fmain\u002F.ai\u002Fskills\u002Fspec-drift-sync\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M7EXE14F7TDGXT5JM4AH0ZVG","shepherd-a-pull-request","Use when reviewing a pull request on GitHub, following up on a review after new pushes, or answering review comments on your own pull request. The `gh` procedure behind the pr-shepherd agent, with draft-then-confirm posting.","https:\u002F\u002Fgithub.com\u002FIntraverseGame\u002Fse-client-kit\u002Fblob\u002Fmain\u002F.ai\u002Fskills\u002Fshepherd-a-pull-request\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M7EXDWWNJK06KPEQNDBVAZT3","prd-to-spec","Use when given a rough PRD, feature request, or raw requirement. Turns it into a spec update, an RFC when the change is significant or contested, and an ADR for hard-to-reverse decisions.","https:\u002F\u002Fgithub.com\u002FIntraverseGame\u002Fse-client-kit\u002Fblob\u002Fmain\u002F.ai\u002Fskills\u002Fprd-to-spec\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M7EXDNP3DKJ987Y40CQNT3D5","add-a-package","Use when adding a new package under packages\u002F (or an app under apps\u002F). Covers the placement check, the dependency direction, public API, wiring, tests, and docs.","https:\u002F\u002Fgithub.com\u002FIntraverseGame\u002Fse-client-kit\u002Fblob\u002Fmain\u002F.ai\u002Fskills\u002Fadd-a-package\u002FSKILL.md",[],[]]