[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M736SS2YXMZ5KQSP9YFENJ6Q":3,"skill-same:01M736SS2YXMZ5KQSP9YFENJ6Q":44},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":13,"repositoryHtmlUrl":26,"repositoryHomepageUrl":6,"ownerHtmlUrl":27,"ownerBlogUrl":6,"ownerDescription":28,"updatedAt":29,"install":30,"resources":43},"01M736SS2YXMZ5KQSP9YFENJ6Q","handoff-session-prompt",null,"Generate a generic next-session initialization prompt from the current conversation context and any user-specified reference files; if a durable record is needed first, write that artifact before the handoff.","graysurf\u002Fagent-runtime-kit","graysurf","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F10785178?v=4","01M736PQHYKHFMNBYF8BT93NM4","01M736PQJWYH7G046WAN7EQCHD",0,"MIT","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fgolden\u002Fcodex\u002Fplugins\u002Fconversation\u002Fskills\u002Fhandoff-session-prompt\u002Fexpected\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: handoff-session-prompt\ndescription: >\n  Generate a generic next-session initialization prompt from the current\n  conversation context and any user-specified reference files; if a durable\n  record is needed first, write that artifact before the handoff.\n---\n\n# Handoff Session Prompt\n\n## Contract\n\nPrereqs:\n\n- User wants a prompt that can initialize a future agent\u002Fsession with enough context to continue work.\n- The prompt should be generic and task-derived, not a hard-coded project template.\n\nInputs:\n\n- User's explicit instruction for the next session, including goal, scope, and any \"do not\" constraints.\n- Current conversation context and conclusions already established in this session.\n- User-specified reference files, URLs, tickets, docs, or commands to include as required reading.\n- Durable project artifacts that already preserve the backlog or decision record, such as plans, runbooks, improvement docs, issues, or\n  tracker links.\n- Optional local repository rules when the handoff concerns a workspace or codebase.\n- Optional current-state anchors when relevant, such as date\u002Ftimezone, workspace path, branch\u002Fstatus, validation already run, and where work\n  stopped.\n\nOutputs:\n\n- A copy-ready session initialization prompt.\n- Optional short notes outside the prompt only when useful, such as assumptions or omitted sensitive details.\n\nExit codes:\n\n- N\u002FA (conversation workflow; no repo scripts)\n\nFailure modes:\n\n- User's requested next-session objective is unclear enough that the handoff would send the next agent in the wrong direction; ask the minimum\n  clarification.\n- Required context or referenced files are unavailable; if the gap could change the objective, safety, or reversibility, ask before\n  producing the prompt. Otherwise, produce a degraded prompt and label the missing context under `Known Gaps` or `Open Questions`.\n- User asks to include secrets, raw credentials, private tokens, or other sensitive operational details; redact or omit them.\n\n## Workflow\n\n1. Identify the future session's job\n   - Extract the concrete objective, expected deliverables, and stopping point.\n   - Preserve the user's requested scope and exclusions.\n   - If the request is underspecified, ask only the minimum clarification needed.\n\n2. Gather task-local context\n   - Use the current conversation conclusions as the primary source.\n   - Read only user-named files or clearly necessary local rules for the target workspace.\n   - If project rules require preflight before file reads or external lookup, follow those rules.\n   - Do not browse or inspect unrelated repositories unless the user asked for that evidence.\n\n3. Prefer durable sources over copying backlogs\n   - If a project doc, plan, issue, ticket, or tracker already holds the durable backlog or decision record, put it under `Read First` and\n     summarize only the facts needed to start safely.\n   - Do not paste an entire durable doc, long backlog, or runbook into the handoff unless the next session will not have access to it or the\n     user explicitly asks for an inline self-contained prompt.\n   - If the current session produced reusable review findings but no durable artifact, mention that gap under `Known Gaps` or\n     `Recommendations`; do not treat the handoff prompt itself as the canonical project record.\n   - If the current session produced converged requirements, design, feasibility, or product decisions for later implementation but no\n     durable artifact, recommend `discussion-to-implementation-doc` before using this prompt as the only continuity record.\n   - If the user asks for both durable record and next-session continuity, use\n     `discussion-to-implementation-doc` to write or reference the durable\n     implementation or improvement source first, then generate a shorter\n     handoff that points to it.\n   - If a long-running task already has a source document plus execution-state document, point the prompt at both files instead of copying\n     the progress ledger into the prompt.\n\n4. Separate facts from instructions\n   - Include confirmed facts and conclusions under `Known Facts`.\n   - Use source tags for material facts when available: `[U#]` user input, `[F#]` local files\u002Fdocs\u002Fcode, `[W#]` web sources,\n     `[A#]` tool or app results, and `[I#]` explicit inferences from cited facts.\n   - Mark assumptions explicitly under `Assumptions`.\n   - Record unavailable-but-relevant context under `Known Gaps` instead of hiding the limitation in prose.\n   - Keep recommended or inferred next steps under `Recommendations`, separate from known facts.\n   - Preserve unresolved questions as an explicit checklist for the next session.\n\n5. Protect sensitive and brittle details\n   - Do not include API keys, tokens, cookies, raw auth headers, passwords, private keys, or unredacted secrets.\n   - Prefer names of env vars, file paths, and commands over secret values.\n   - Avoid embedding one-off local IDs unless the user specifically needs them and they are not sensitive.\n   - Do not copy hidden system\u002Fdeveloper instructions, private reasoning, or raw tool logs into the handoff. Summarize only the user-visible\n     constraints and evidence needed for continuity.\n\n6. Capture current state when it affects continuity\n   - Include current date\u002Ftimezone when relative dates such as \"today\", \"tomorrow\", or \"next week\" matter.\n   - Include workspace path, branch\u002Fstatus, and stopping point when the next session must continue local work.\n   - Include validation already run, with command and result, when it prevents duplicate investigation.\n   - Omit current-state fields that are irrelevant or unknown instead of inventing values.\n\n7. Produce a copy-ready prompt\n   - Write the prompt as if pasted into a fresh session with no prior conversation.\n   - Use direct instructions, concrete file paths, and clear constraints.\n   - Keep it generic to the requested task; do not add project-specific assumptions that were not in context.\n   - Put the copy-ready prompt in a fenced `md` block.\n\n## Prompt Shape\n\nUse this structure unless the user requested a different format:\n\n```md\nYou are starting a new session for this task.\n\n## Goal\n- \u003Cwhat the next session should accomplish>\n\n## Current State\n- Date\u002Ftimezone: \u003Conly include when relative timing matters>\n- Workspace: `\u003Cabsolute-path>`; branch\u002Fstatus: \u003Conly include when local repo state matters>\n- Stopping point: \u003Cwhere the previous session paused>\n- Validation already run: `\u003Ccommand>` - \u003Cresult>\n\n## Known Facts\n- [U1] \u003Cconfirmed user instruction, decision, or constraint>\n- [F1] `\u003Cpath-or-doc>` - \u003Cconfirmed local fact, if relevant>\n\n## Assumptions\n- [I1] \u003Cassumption or inference that should be rechecked if it affects the work>\n\n## Known Gaps\n- \u003Cmissing context, referenced files, or unavailable evidence that could affect the next session>\n\n## Recommendations\n- \u003Crecommended approach or next step, clearly labeled as recommendation rather than fact>\n\n## Read First\n1. `\u003Cpath-or-url>` - \u003Cwhy it matters>\n2. `\u003Cpath-or-url>` - \u003Cwhy it matters>\n\n## Task\n1. \u003Cfirst concrete action>\n2. \u003Csecond concrete action>\n3. \u003Cverification or reporting action>\n\n## Constraints\n- Do not \u003Cforbidden action>.\n- Keep \u003Cscope\u002Fstyle\u002Fsecurity rule>.\n- Redact or omit secrets and sensitive operational details.\n\n## Expected Output\n- \u003Cartifact or response the next session should produce>\n- \u003Csummary\u002Fevidence\u002Fchecks to report>\n\n## Open Questions\n- \u003Conly unresolved items that should block or influence the work>\n```\n\n## Plan-Tracking Handoff\n\nWhen the next session resumes a plan-tracking issue governed by\n`deliver-plan-tracking-issue`, include a dedicated \"Plan-tracking\nhandoff\" subsection inside `Known Facts` (or its own clearly labeled\nblock) that dumps four concrete elements so the next agent does not have\nto re-derive run-state context:\n\n1. **Current `run-state.json` contents** — the typed run state for this\n   issue. Either paste the JSON verbatim or reference the path\n   (`\u003CAGENT_HOME>\u002Fout\u002F...\u002Frun-state.json`) and summarize `phase`,\n   `selected_task`, `branch`, `linked_pr`, and any `validation_*`\n   fields that are non-empty.\n2. **Last N entries of `events.jsonl`** — default N=5. Each event line\n   is the source of truth for what happened since the previous\n   checkpoint. Trim to the most recent N when the file is long, and\n   note `last_n: \u003CN>` so the next session can read more if needed.\n3. **Open ledger rows** — every row in the bundle's\n   `\u003Cslug>-execution-state.md` `## Task Ledger` table whose\n   `Status ∈ {pending, in-progress}`. Format as `\u003Ctask_id> | \u003Cstatus>\n   | \u003Cevidence-or-empty>` so the next agent knows exactly which\n   `plan-tooling ledger-update --task \u003Cid> --status done --evidence\n   \u003Cevidence>` calls are still owed.\n4. **Latest linked PR refs + `state` lifecycle comment URL** — the\n   `linked_pr` value(s) from the run state and the URL of the most\n   recent `state` role comment on the provider issue (via\n   `gh issue view \u003Cissue> --json comments` or the controller's\n   reconciled evidence). This lets the next session reconcile run\n   state against provider truth without re-running every gate.\n\nKeep the dump compact: the goal is the next session can call\n`plan-issue tracking status --expect-visible` once and confirm the\nhandoff matches reality, not re-discover the entire history.\n\n## Quality Bar\n\n- The prompt must be understandable without access to the old conversation.\n- The prompt must tell the next session what to read, what to do, what not to do, and what to report.\n- Material context must be traceable as facts, assumptions, recommendations, open questions, or current-state anchors.\n- The prompt should not contain hidden implementation decisions; if an approach is a recommendation, label it as such.\n- The prompt should point to durable project records instead of becoming the only copy of a backlog or decision record.\n- The prompt should be concise enough to paste comfortably, but complete enough to prevent predictable re-discovery.\n","# Handoff Session Prompt\n\n## Contract\n\nPrereqs:\n\n- User wants a prompt that can initialize a future agent\u002Fsession with enough context to continue work.\n- The prompt should be generic and task-derived, not a hard-coded project template.\n\nInputs:\n\n- User's explicit instruction for the next session, including goal, scope, and any \"do not\" constraints.\n- Current conversation context and conclusions already established in this session.\n- User-specified reference files, URLs, tickets, docs, or commands to include as required reading.\n- Durable project artifacts that already preserve the backlog or decision record, such as plans, runbooks, improvement docs, issues, or\n  tracker links.\n- Optional local repository rules when the handoff concerns a workspace or codebase.\n- Optional current-state anchors when relevant, such as date\u002Ftimezone, workspace path, branch\u002Fstatus, validation already run, and where work\n  stopped.\n\nOutputs:\n\n- A copy-ready session initialization prompt.\n- Optional short notes outside the prompt only when useful, such as assumptions or omitted sensitive details.\n\nExit codes:\n\n- N\u002FA (conversation workflow; no repo scripts)\n\nFailure modes:\n\n- User's requested next-session objective is unclear enough that the handoff would send the next agent in the wrong direction; ask the minimum\n  clarification.\n- Required context or referenced files are unavailable; if the gap could change the objective, safety, or reversibility, ask before\n  producing the prompt. Otherwise, produce a degraded prompt and label the missing context under `Known Gaps` or `Open Questions`.\n- User asks to include secrets, raw credentials, private tokens, or other sensitive operational details; redact or omit them.\n\n## Workflow\n\n1. Identify the future session's job\n   - Extract the concrete objective, expected deliverables, and stopping point.\n   - Preserve the user's requested scope and exclusions.\n   - If the request is underspecified, ask only the minimum clarification needed.\n\n2. Gather task-local context\n   - Use the current conversation conclusions as the primary source.\n   - Read only user-named files or clearly necessary local rules for the target workspace.\n   - If project rules require preflight before file reads or external lookup, follow those rules.\n   - Do not browse or inspect unrelated repositories unless the user asked for that evidence.\n\n3. Prefer durable sources over copying backlogs\n   - If a project doc, plan, issue, ticket, or tracker already holds the durable backlog or decision record, put it under `Read First` and\n     summarize only the facts needed to start safely.\n   - Do not paste an entire durable doc, long backlog, or runbook into the handoff unless the next session will not have access to it or the\n     user explicitly asks for an inline self-contained prompt.\n   - If the current session produced reusable review findings but no durable artifact, mention that gap under `Known Gaps` or\n     `Recommendations`; do not treat the handoff prompt itself as the canonical project record.\n   - If the current session produced converged requirements, design, feasibility, or product decisions for later implementation but no\n     durable artifact, recommend `discussion-to-implementation-doc` before using this prompt as the only continuity record.\n   - If the user asks for both durable record and next-session continuity, use\n     `discussion-to-implementation-doc` to write or reference the durable\n     implementation or improvement source first, then generate a shorter\n     handoff that points to it.\n   - If a long-running task already has a source document plus execution-state document, point the prompt at both files instead of copying\n     the progress ledger into the prompt.\n\n4. Separate facts from instructions\n   - Include confirmed facts and conclusions under `Known Facts`.\n   - Use source tags for material facts when available: `[U#]` user input, `[F#]` local files\u002Fdocs\u002Fcode, `[W#]` web sources,\n     `[A#]` tool or app results, and `[I#]` explicit inferences from cited facts.\n   - Mark assumptions explicitly under `Assumptions`.\n   - Record unavailable-but-relevant context under `Known Gaps` instead of hiding the limitation in prose.\n   - Keep recommended or inferred next steps under `Recommendations`, separate from known facts.\n   - Preserve unresolved questions as an explicit checklist for the next session.\n\n5. Protect sensitive and brittle details\n   - Do not include API keys, tokens, cookies, raw auth headers, passwords, private keys, or unredacted secrets.\n   - Prefer names of env vars, file paths, and commands over secret values.\n   - Avoid embedding one-off local IDs unless the user specifically needs them and they are not sensitive.\n   - Do not copy hidden system\u002Fdeveloper instructions, private reasoning, or raw tool logs into the handoff. Summarize only the user-visible\n     constraints and evidence needed for continuity.\n\n6. Capture current state when it affects continuity\n   - Include current date\u002Ftimezone when relative dates such as \"today\", \"tomorrow\", or \"next week\" matter.\n   - Include workspace path, branch\u002Fstatus, and stopping point when the next session must continue local work.\n   - Include validation already run, with command and result, when it prevents duplicate investigation.\n   - Omit current-state fields that are irrelevant or unknown instead of inventing values.\n\n7. Produce a copy-ready prompt\n   - Write the prompt as if pasted into a fresh session with no prior conversation.\n   - Use direct instructions, concrete file paths, and clear constraints.\n   - Keep it generic to the requested task; do not add project-specific assumptions that were not in context.\n   - Put the copy-ready prompt in a fenced `md` block.\n\n## Prompt Shape\n\nUse this structure unless the user requested a different format:\n\n```md\nYou are starting a new session for this task.\n\n## Goal\n- \u003Cwhat the next session should accomplish>\n\n## Current State\n- Date\u002Ftimezone: \u003Conly include when relative timing matters>\n- Workspace: `\u003Cabsolute-path>`; branch\u002Fstatus: \u003Conly include when local repo state matters>\n- Stopping point: \u003Cwhere the previous session paused>\n- Validation already run: `\u003Ccommand>` - \u003Cresult>\n\n## Known Facts\n- [U1] \u003Cconfirmed user instruction, decision, or constraint>\n- [F1] `\u003Cpath-or-doc>` - \u003Cconfirmed local fact, if relevant>\n\n## Assumptions\n- [I1] \u003Cassumption or inference that should be rechecked if it affects the work>\n\n## Known Gaps\n- \u003Cmissing context, referenced files, or unavailable evidence that could affect the next session>\n\n## Recommendations\n- \u003Crecommended approach or next step, clearly labeled as recommendation rather than fact>\n\n## Read First\n1. `\u003Cpath-or-url>` - \u003Cwhy it matters>\n2. `\u003Cpath-or-url>` - \u003Cwhy it matters>\n\n## Task\n1. \u003Cfirst concrete action>\n2. \u003Csecond concrete action>\n3. \u003Cverification or reporting action>\n\n## Constraints\n- Do not \u003Cforbidden action>.\n- Keep \u003Cscope\u002Fstyle\u002Fsecurity rule>.\n- Redact or omit secrets and sensitive operational details.\n\n## Expected Output\n- \u003Cartifact or response the next session should produce>\n- \u003Csummary\u002Fevidence\u002Fchecks to report>\n\n## Open Questions\n- \u003Conly unresolved items that should block or influence the work>\n```\n\n## Plan-Tracking Handoff\n\nWhen the next session resumes a plan-tracking issue governed by\n`deliver-plan-tracking-issue`, include a dedicated \"Plan-tracking\nhandoff\" subsection inside `Known Facts` (or its own clearly labeled\nblock) that dumps four concrete elements so the next agent does not have\nto re-derive run-state context:\n\n1. **Current `run-state.json` contents** — the typed run state for this\n   issue. Either paste the JSON verbatim or reference the path\n   (`\u003CAGENT_HOME>\u002Fout\u002F...\u002Frun-state.json`) and summarize `phase`,\n   `selected_task`, `branch`, `linked_pr`, and any `validation_*`\n   fields that are non-empty.\n2. **Last N entries of `events.jsonl`** — default N=5. Each event line\n   is the source of truth for what happened since the previous\n   checkpoint. Trim to the most recent N when the file is long, and\n   note `last_n: \u003CN>` so the next session can read more if needed.\n3. **Open ledger rows** — every row in the bundle's\n   `\u003Cslug>-execution-state.md` `## Task Ledger` table whose\n   `Status ∈ {pending, in-progress}`. Format as `\u003Ctask_id> | \u003Cstatus>\n   | \u003Cevidence-or-empty>` so the next agent knows exactly which\n   `plan-tooling ledger-update --task \u003Cid> --status done --evidence\n   \u003Cevidence>` calls are still owed.\n4. **Latest linked PR refs + `state` lifecycle comment URL** — the\n   `linked_pr` value(s) from the run state and the URL of the most\n   recent `state` role comment on the provider issue (via\n   `gh issue view \u003Cissue> --json comments` or the controller's\n   reconciled evidence). This lets the next session reconcile run\n   state against provider truth without re-running every gate.\n\nKeep the dump compact: the goal is the next session can call\n`plan-issue tracking status --expect-visible` once and confirm the\nhandoff matches reality, not re-discover the entire history.\n\n## Quality Bar\n\n- The prompt must be understandable without access to the old conversation.\n- The prompt must tell the next session what to read, what to do, what not to do, and what to report.\n- Material context must be traceable as facts, assumptions, recommendations, open questions, or current-state anchors.\n- The prompt should not contain hidden implementation decisions; if an approach is a recommendation, label it as such.\n- The prompt should point to durable project records instead of becoming the only copy of a backlog or decision record.\n- The prompt should be concise enough to paste comfortably, but complete enough to prevent predictable re-discovery.\n",{"name":5,"description":7},"main","da7119f2ebece510d6955cea72743069f5496c73","tests\u002Fgolden\u002Fcodex\u002Fplugins\u002Fconversation\u002Fskills\u002Fhandoff-session-prompt\u002Fexpected","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit","https:\u002F\u002Fgithub.com\u002Fgraysurf","Building reliable tools for teams of AI coding agents · Rust \u002F TypeScript · @sympoies","2026-08-29 05:27:47",{"source":8,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add graysurf\u002Fagent-runtime-kit --skill handoff-session-prompt",{"tool":36,"command":37},"bunx","bunx skills add graysurf\u002Fagent-runtime-kit --skill handoff-session-prompt",{"tool":39,"command":40},"pnpm","pnpm dlx skills add graysurf\u002Fagent-runtime-kit --skill handoff-session-prompt","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fgolden\u002Fcodex\u002Fplugins\u002Fconversation\u002Fskills\u002Fhandoff-session-prompt\u002Fexpected\u002FSKILL.md",true,[],[45,51,56,63,70,77],{"skillNo":46,"name":47,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":48,"categories":49,"keywords":50,"publishedAt":18},"01M736XWK14DFRERMQ8RT3HS0Z","removable-skill","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fruntime-smoke\u002Ffixtures\u002Fskill-lifecycle\u002Fremove-skill\u002Ftests\u002Fgolden\u002Fcodex\u002Fplugins\u002Ffixture\u002Fskills\u002Fremovable-skill\u002FSKILL.md",[],[],{"skillNo":52,"name":47,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":53,"categories":54,"keywords":55,"publishedAt":18},"01M736XTPQEKAYESDM1REN2KAN","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fruntime-smoke\u002Ffixtures\u002Fskill-lifecycle\u002Fremove-skill\u002Ftests\u002Fgolden\u002Fclaude\u002Fplugins\u002Ffixture\u002Fskills\u002Fremovable-skill\u002FSKILL.md",[],[],{"skillNo":57,"name":58,"nameZh":6,"description":59,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":60,"categories":61,"keywords":62,"publishedAt":18},"01M736XRMWHMJB4X563G9J6QYY","removable-project-skill","Fixture project-local skill used to prove remove-project-skill dry-run coverage.","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fruntime-smoke\u002Ffixtures\u002Fskill-lifecycle\u002Fremove-project-skill\u002F.agents\u002Fskills\u002Fremovable-project-skill\u002FSKILL.md",[],[],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":18},"01M736XPT65YZMPCBD4K3RHV4P","project-sample-skill","Fixture project-local skill used to prove create-project-skill governance coverage.","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fruntime-smoke\u002Ffixtures\u002Fskill-lifecycle\u002Fcreate-project-skill\u002F.agents\u002Fskills\u002Fproject-sample-skill\u002FSKILL.md",[],[],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":18},"01M736XKZVYWVQA1YB0P459WND","project-local-skill","Fixture project-local skill used by runtime-kit project-local smoke checks.","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fprojects\u002Fproject-local-smoke\u002F.agents\u002Fskills\u002Fproject-local-skill\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":18},"01M736XHA73KZ5G2TKWSED5P1B","topic-radar","Aggregate read-only AI and technology trend signals (Polymarket, Hacker News, GitHub, arXiv, Hugging Face, news search) into source-grounded Markdown or JSON; the lower-level data engine behind `daily-brief`.","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fgolden\u002Fhermes\u002Fplugins\u002Freporting\u002Fskills\u002Ftopic-radar\u002Fexpected\u002FSKILL.md",[],[]]