[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7F66HAT0RG9TMXQ28GJXERR":3,"skill-same:01M7F66HAT0RG9TMXQ28GJXERR":44},{"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":14,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":29,"install":30,"resources":43},"01M7F66HAT0RG9TMXQ28GJXERR","agents-and-skills-authoring",null,"Construct overlay on the official agent-development, skill-development, and skill-creator skills. Current frontmatter rules, what plugin agents ignore, startup context, model precedence, nesting, skill compaction, `!` dynamic context, and workflows, with corrections where the official skills are out of date. Use together with them before adding or editing a construct agent, skill, role, or workflow, or when an agent or skill behaves differently than its file says.","carsimax\u002Fconstruct","carsimax","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F42248429?v=4","01M7F6660B38RMTKRYJ2DQM39T","Maximiliano Carsi","01M7F6661MDPBN2351WPZNV6CT",1,"MIT","https:\u002F\u002Fgithub.com\u002Fcarsimax\u002Fconstruct\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fagents-and-skills-authoring\u002FSKILL.md",[],[],"2026-11-16 10:05:34",false,"---\nname: agents-and-skills-authoring\ndescription: Construct overlay on the official agent-development, skill-development, and skill-creator skills. Current frontmatter rules, what plugin agents ignore, startup context, model precedence, nesting, skill compaction, `!` dynamic context, and workflows, with corrections where the official skills are out of date. Use together with them before adding or editing a construct agent, skill, role, or workflow, or when an agent or skill behaves differently than its file says.\n---\n\n# Agents, skills, and workflows in construct\n\n**Load the official skills first**: `agent-development` (system prompt design, triggering examples, `scripts\u002Fvalidate-agent.sh`), `skill-development` (structure, progressive disclosure), and `skill-creator` (anthropics\u002Fskills: drafting, evaluating, and tuning a skill's description with its eval loop). This file adds what they lack and corrects what is stale. Order of authority: the live docs, then this file, then the official skills.\n\n## Corrections to the official skills\n\n| Official skill says | Current docs |\n| :- | :- |\n| agent `model` and `color` are required | only `name` and `description` are required |\n| agent `model`: `inherit`, `sonnet`, `opus`, `haiku` | also `fable` and full model ids |\n| `color`: `blue`, `cyan`, `green`, `yellow`, `magenta`, `red` | `red`, `blue`, `green`, `yellow`, `purple`, `orange`, `pink`, `cyan` |\n| no `effort`, `maxTurns`, `skills`, `memory`, `isolation`, `disallowedTools`, ... | all exist; plugin agents ignore `permissionMode`, `hooks`, `mcpServers`, `initialPrompt` |\n| skill frontmatter mostly `name`, `description`, `version` | `version` is not a Claude Code field; see the field list below |\n| `skill-creator` eval scripts (`scripts\u002Frun_eval.py`) | they run `claude -p` and Python; for construct's plugin skills, prefer `claude plugin eval` (see `plugin-evals`) |\n\nSources (read 2026-10-23): [sub-agents](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fsub-agents.md), [skills](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fskills.md), [workflows](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fworkflows.md). Re-read before relying on a version-gated detail. Lines marked **observed** come from our own live probes (construct memory), not the docs.\n\n## Subagent files (`plugins\u002Fconstruct\u002Fagents\u002F*.md`)\n\n- Required: `name` (≤ 256 chars, no `:`, hooks see it as `agent_type`, scoped as `construct:\u003Cname>`) and `description`. Unknown fields are **silently ignored**; `claude --debug` logs `does not recognize or apply` (v2.1.296+).\n- **Plugin agents ignore `permissionMode`, `hooks`, `mcpServers`, `initialPrompt`.** Guards go in `hooks\u002Fhooks.json`, never in agent frontmatter.\n- Honored: `tools` (omitted = inherit all), `disallowedTools` (applied first; `Bash(git push *)` removes the whole tool), `model` (`sonnet`, `opus`, `haiku`, `fable`, full id, `inherit`), `effort` (`low`…`max`), `maxTurns` (partial result, resumable), `autoCompactWindow` (integer `100000`–`1000000`; `200k` reads as 200), `skills` (preloads full content, first 32; cannot preload `disable-model-invocation` skills; does not restrict access), `memory` (`user`\u002F`project`\u002F`local` `MEMORY.md`, first 200 lines or 25 KB, auto-enables Read\u002FWrite\u002FEdit), `background`, `omitClaudeMd` (v2.1.271+), `isolation: worktree` (branches from the **default branch**, not the parent's `HEAD`), `color`, `experimental.cacheTtl` (`5m`\u002F`1h`).\n- Unparseable YAML: the agent still loads under its file name with every field ignored. A JSON-only lens with broken frontmatter silently loses its `tools` limits; `claude plugin validate --strict` must stay green.\n- Scope priority: managed > `--agents` > `.claude\u002Fagents\u002F` > `~\u002F.claude\u002Fagents\u002F` > plugin (lowest). A user agent with the same name shadows ours.\n\n### Model precedence (routing)\n\n1. Per-invocation `model` on the Agent call; **a mod's `agent.spawn` hook replaces it** (that is how `hooks\u002Frouting.js` routes).\n2. Frontmatter `model` (construct's fallback when the mod cannot run).\n3. `CLAUDE_CODE_SUBAGENT_MODEL`.\n4. The main model.\n\n`CLAUDE_CODE_SUBAGENT_MODEL_FORCE=1` (v2.1.257+) ignores 1 and 2. A family alias resolves to the main model's exact id when it is in that family. Values are checked against `availableModels`. Frontmatter `effort` overrides session effort but not `CLAUDE_CODE_EFFORT_LEVEL`; Agent tool `effort` exists from v2.1.292.\n\n### What a subagent sees\n\n- Gets: its own body as system prompt (not Claude Code's), the task message, all CLAUDE.md levels and AGENTS.md (Explore and Plan skip them), a git status snapshot, preloaded skills, the `SubagentStart` `additionalContext` (our subagent contract).\n- Does **not** get: parent history, skills the parent loaded, files the parent read, the output style, parent auto memory. Hand off by reference to `.vault\u002Ftasks\u002F\u003Cfeature>.md`, never by \"as discussed\".\n- Combined non-built-in descriptions over 15,000 tokens raise a startup warning. Keep `description` short and the detail in the body.\n\n### Nesting, concurrency, background\n\n- Nesting depth 3 below main by default (`CLAUDE_CODE_MAX_SUBAGENT_SPAWN_DEPTH`; `1` disables). A subagent spawns only if `Agent` is in its tools. Construct's rule \"subagents never launch agents\" therefore needs `Agent` absent from every role's `tools`.\n- 20 concurrent subagents (`CLAUDE_CODE_MAX_CONCURRENT_SUBAGENTS`), error `Concurrent subagent limit reached`.\n- Subagents spawned from main start in the background by default. Background subagents get a narrower built-in tool set; every subagent loses `AskUserQuestion`, `EnterPlanMode`, `Workflow`, `ScheduleWakeup`. A worker cannot ask the user directly, hence the `needs_interaction` relay.\n- Resume with `SendMessage` (`to` = id or name) keeps history and tools. Transcripts: `~\u002F.claude\u002Fprojects\u002F\u003Cproject>\u002F\u003Csession>\u002Fsubagents\u002Fagent-\u003Cid>.jsonl`.\n\n## Skill files (`plugins\u002Fconstruct\u002Fskills\u002F*\u002FSKILL.md`)\n\n- Frontmatter is read only if `---` is the first line. Invalid YAML: the skill loads with **no fields**, so it cannot be matched by description. Unknown fields are ignored.\n- Fields: `name` (sets the last segment, `\u002Fconstruct:\u003Cname>`), `description` + `when_to_use` (truncated together at **1,536 chars**; key use case first), `disable-model-invocation` (also hides the description and blocks preload), `user-invocable`, `argument-hint`, `arguments` (`$name`), `allowed-tools` (pre-approves for the invoking turn only; does not restrict), `disallowed-tools`, `model` and `effort` (rest of the turn), `context: fork` + `agent` + `background`, `hooks` (registered on invocation, stay for the session), `paths` (globs relative to cwd), `shell`.\n- Substitutions: `$ARGUMENTS`, `$N`, `${CLAUDE_SESSION_ID}`, `${CLAUDE_EFFORT}`, `${CLAUDE_PROJECT_DIR}`, `${CLAUDE_SKILL_DIR}` (the skill's own directory, not the plugin root), `${CLAUDE_PLUGIN_ROOT}`, `${CLAUDE_PLUGIN_DATA}`, `${user_config.KEY}` (non-sensitive).\n- Loading: descriptions are always in context; the body loads on invocation as one message and is not re-read. Listing budget is 1% of the context window; overflow drops the least-invoked descriptions first.\n- **Compaction keeps only the first 5,000 tokens** of the latest invocation of each skill, 25,000 tokens total, most recent first. Put the rules that must survive at the top. Keep `SKILL.md` under 500 lines.\n- Plugin skills coexist with same-named skills elsewhere and are not affected by `skillOverrides`.\n\n### `!` dynamic context (`skill-context.mjs`)\n\n- `` !`cmd` `` at line start or after whitespace, or a ` ```! ` fence. Runs once before Claude sees the skill; output is not re-scanned; cwd is the session shell's cwd (use `${CLAUDE_SKILL_DIR}` or `${CLAUDE_PLUGIN_ROOT}` for paths); stderr is merged; 2-minute timeout.\n- **Any non-zero exit aborts the whole invocation** (`Shell command failed for pattern \"...\"`), except exit 1 from search\u002Fcompare commands. **Observed** in our probes too. That is why `skill-context.mjs` always exits 0 and prints `(state unavailable: \u003Creason>)`.\n- Injected commands never prompt: outside auto mode anything but allow aborts (`Shell command permission check failed`). Pre-approve through `allowed-tools`.\n- `disableSkillShellExecution: true` replaces the command with `[shell command execution disabled by policy]` for plugin skills too. The skill text must still make sense without the injected state.\n\n## Workflows (`workflows\u002F*.js`)\n\n- `export const meta = { name, description }` must be the first statement, a plain literal. Plugin workflows run as `\u002Fconstruct:\u003Cname>`.\n- API: `agent(prompt, { schema, label, stallMs })` (resolves `null` on stop or unrecoverable error), `parallel()`, `pipeline()`, `phase()`, `log()`, global `args`. No `import()`, filesystem, shell, `Date.now()`, or `Math.random()`.\n- Limits: 16 concurrent agents by default, 1,000 agents per run, 4,096 items per `parallel`\u002F`pipeline`.\n- Hooks for workflow agents are not documented. **Observed**: `agent()` in a workflow does **not** fire `PreToolUse:Agent`, so `review-context.mjs` and delegation admission do not see those launches. Never route a gate-relevant launch (lens, writer) through a workflow.\n- In `-p` the launch needs an allow rule (`Workflow(\u003Cname>)`), auto or bypass mode, or a `PreToolUse` allow. Disable with `disableWorkflows` or `CLAUDE_CODE_DISABLE_WORKFLOWS=1`.\n\n## Checks for a change here\n\n`node --test 'tests\u002F**\u002F*.test.mjs'` (agents and skills tests), `claude plugin validate --strict .\u002Fplugins\u002Fconstruct`, and, for delegation behavior, an eval case (see `plugin-evals`).\n","# Agents, skills, and workflows in construct\n\n**Load the official skills first**: `agent-development` (system prompt design, triggering examples, `scripts\u002Fvalidate-agent.sh`), `skill-development` (structure, progressive disclosure), and `skill-creator` (anthropics\u002Fskills: drafting, evaluating, and tuning a skill's description with its eval loop). This file adds what they lack and corrects what is stale. Order of authority: the live docs, then this file, then the official skills.\n\n## Corrections to the official skills\n\n| Official skill says | Current docs |\n| :- | :- |\n| agent `model` and `color` are required | only `name` and `description` are required |\n| agent `model`: `inherit`, `sonnet`, `opus`, `haiku` | also `fable` and full model ids |\n| `color`: `blue`, `cyan`, `green`, `yellow`, `magenta`, `red` | `red`, `blue`, `green`, `yellow`, `purple`, `orange`, `pink`, `cyan` |\n| no `effort`, `maxTurns`, `skills`, `memory`, `isolation`, `disallowedTools`, ... | all exist; plugin agents ignore `permissionMode`, `hooks`, `mcpServers`, `initialPrompt` |\n| skill frontmatter mostly `name`, `description`, `version` | `version` is not a Claude Code field; see the field list below |\n| `skill-creator` eval scripts (`scripts\u002Frun_eval.py`) | they run `claude -p` and Python; for construct's plugin skills, prefer `claude plugin eval` (see `plugin-evals`) |\n\nSources (read 2026-10-23): [sub-agents](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fsub-agents.md), [skills](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fskills.md), [workflows](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fworkflows.md). Re-read before relying on a version-gated detail. Lines marked **observed** come from our own live probes (construct memory), not the docs.\n\n## Subagent files (`plugins\u002Fconstruct\u002Fagents\u002F*.md`)\n\n- Required: `name` (≤ 256 chars, no `:`, hooks see it as `agent_type`, scoped as `construct:\u003Cname>`) and `description`. Unknown fields are **silently ignored**; `claude --debug` logs `does not recognize or apply` (v2.1.296+).\n- **Plugin agents ignore `permissionMode`, `hooks`, `mcpServers`, `initialPrompt`.** Guards go in `hooks\u002Fhooks.json`, never in agent frontmatter.\n- Honored: `tools` (omitted = inherit all), `disallowedTools` (applied first; `Bash(git push *)` removes the whole tool), `model` (`sonnet`, `opus`, `haiku`, `fable`, full id, `inherit`), `effort` (`low`…`max`), `maxTurns` (partial result, resumable), `autoCompactWindow` (integer `100000`–`1000000`; `200k` reads as 200), `skills` (preloads full content, first 32; cannot preload `disable-model-invocation` skills; does not restrict access), `memory` (`user`\u002F`project`\u002F`local` `MEMORY.md`, first 200 lines or 25 KB, auto-enables Read\u002FWrite\u002FEdit), `background`, `omitClaudeMd` (v2.1.271+), `isolation: worktree` (branches from the **default branch**, not the parent's `HEAD`), `color`, `experimental.cacheTtl` (`5m`\u002F`1h`).\n- Unparseable YAML: the agent still loads under its file name with every field ignored. A JSON-only lens with broken frontmatter silently loses its `tools` limits; `claude plugin validate --strict` must stay green.\n- Scope priority: managed > `--agents` > `.claude\u002Fagents\u002F` > `~\u002F.claude\u002Fagents\u002F` > plugin (lowest). A user agent with the same name shadows ours.\n\n### Model precedence (routing)\n\n1. Per-invocation `model` on the Agent call; **a mod's `agent.spawn` hook replaces it** (that is how `hooks\u002Frouting.js` routes).\n2. Frontmatter `model` (construct's fallback when the mod cannot run).\n3. `CLAUDE_CODE_SUBAGENT_MODEL`.\n4. The main model.\n\n`CLAUDE_CODE_SUBAGENT_MODEL_FORCE=1` (v2.1.257+) ignores 1 and 2. A family alias resolves to the main model's exact id when it is in that family. Values are checked against `availableModels`. Frontmatter `effort` overrides session effort but not `CLAUDE_CODE_EFFORT_LEVEL`; Agent tool `effort` exists from v2.1.292.\n\n### What a subagent sees\n\n- Gets: its own body as system prompt (not Claude Code's), the task message, all CLAUDE.md levels and AGENTS.md (Explore and Plan skip them), a git status snapshot, preloaded skills, the `SubagentStart` `additionalContext` (our subagent contract).\n- Does **not** get: parent history, skills the parent loaded, files the parent read, the output style, parent auto memory. Hand off by reference to `.vault\u002Ftasks\u002F\u003Cfeature>.md`, never by \"as discussed\".\n- Combined non-built-in descriptions over 15,000 tokens raise a startup warning. Keep `description` short and the detail in the body.\n\n### Nesting, concurrency, background\n\n- Nesting depth 3 below main by default (`CLAUDE_CODE_MAX_SUBAGENT_SPAWN_DEPTH`; `1` disables). A subagent spawns only if `Agent` is in its tools. Construct's rule \"subagents never launch agents\" therefore needs `Agent` absent from every role's `tools`.\n- 20 concurrent subagents (`CLAUDE_CODE_MAX_CONCURRENT_SUBAGENTS`), error `Concurrent subagent limit reached`.\n- Subagents spawned from main start in the background by default. Background subagents get a narrower built-in tool set; every subagent loses `AskUserQuestion`, `EnterPlanMode`, `Workflow`, `ScheduleWakeup`. A worker cannot ask the user directly, hence the `needs_interaction` relay.\n- Resume with `SendMessage` (`to` = id or name) keeps history and tools. Transcripts: `~\u002F.claude\u002Fprojects\u002F\u003Cproject>\u002F\u003Csession>\u002Fsubagents\u002Fagent-\u003Cid>.jsonl`.\n\n## Skill files (`plugins\u002Fconstruct\u002Fskills\u002F*\u002FSKILL.md`)\n\n- Frontmatter is read only if `---` is the first line. Invalid YAML: the skill loads with **no fields**, so it cannot be matched by description. Unknown fields are ignored.\n- Fields: `name` (sets the last segment, `\u002Fconstruct:\u003Cname>`), `description` + `when_to_use` (truncated together at **1,536 chars**; key use case first), `disable-model-invocation` (also hides the description and blocks preload), `user-invocable`, `argument-hint`, `arguments` (`$name`), `allowed-tools` (pre-approves for the invoking turn only; does not restrict), `disallowed-tools`, `model` and `effort` (rest of the turn), `context: fork` + `agent` + `background`, `hooks` (registered on invocation, stay for the session), `paths` (globs relative to cwd), `shell`.\n- Substitutions: `$ARGUMENTS`, `$N`, `${CLAUDE_SESSION_ID}`, `${CLAUDE_EFFORT}`, `${CLAUDE_PROJECT_DIR}`, `${CLAUDE_SKILL_DIR}` (the skill's own directory, not the plugin root), `${CLAUDE_PLUGIN_ROOT}`, `${CLAUDE_PLUGIN_DATA}`, `${user_config.KEY}` (non-sensitive).\n- Loading: descriptions are always in context; the body loads on invocation as one message and is not re-read. Listing budget is 1% of the context window; overflow drops the least-invoked descriptions first.\n- **Compaction keeps only the first 5,000 tokens** of the latest invocation of each skill, 25,000 tokens total, most recent first. Put the rules that must survive at the top. Keep `SKILL.md` under 500 lines.\n- Plugin skills coexist with same-named skills elsewhere and are not affected by `skillOverrides`.\n\n### `!` dynamic context (`skill-context.mjs`)\n\n- `` !`cmd` `` at line start or after whitespace, or a ` ```! ` fence. Runs once before Claude sees the skill; output is not re-scanned; cwd is the session shell's cwd (use `${CLAUDE_SKILL_DIR}` or `${CLAUDE_PLUGIN_ROOT}` for paths); stderr is merged; 2-minute timeout.\n- **Any non-zero exit aborts the whole invocation** (`Shell command failed for pattern \"...\"`), except exit 1 from search\u002Fcompare commands. **Observed** in our probes too. That is why `skill-context.mjs` always exits 0 and prints `(state unavailable: \u003Creason>)`.\n- Injected commands never prompt: outside auto mode anything but allow aborts (`Shell command permission check failed`). Pre-approve through `allowed-tools`.\n- `disableSkillShellExecution: true` replaces the command with `[shell command execution disabled by policy]` for plugin skills too. The skill text must still make sense without the injected state.\n\n## Workflows (`workflows\u002F*.js`)\n\n- `export const meta = { name, description }` must be the first statement, a plain literal. Plugin workflows run as `\u002Fconstruct:\u003Cname>`.\n- API: `agent(prompt, { schema, label, stallMs })` (resolves `null` on stop or unrecoverable error), `parallel()`, `pipeline()`, `phase()`, `log()`, global `args`. No `import()`, filesystem, shell, `Date.now()`, or `Math.random()`.\n- Limits: 16 concurrent agents by default, 1,000 agents per run, 4,096 items per `parallel`\u002F`pipeline`.\n- Hooks for workflow agents are not documented. **Observed**: `agent()` in a workflow does **not** fire `PreToolUse:Agent`, so `review-context.mjs` and delegation admission do not see those launches. Never route a gate-relevant launch (lens, writer) through a workflow.\n- In `-p` the launch needs an allow rule (`Workflow(\u003Cname>)`), auto or bypass mode, or a `PreToolUse` allow. Disable with `disableWorkflows` or `CLAUDE_CODE_DISABLE_WORKFLOWS=1`.\n\n## Checks for a change here\n\n`node --test 'tests\u002F**\u002F*.test.mjs'` (agents and skills tests), `claude plugin validate --strict .\u002Fplugins\u002Fconstruct`, and, for delegation behavior, an eval case (see `plugin-evals`).\n",{"name":5,"description":7},"main","22badcb55851e82eb81a14d18204460bd3881d4a",".claude\u002Fskills\u002Fagents-and-skills-authoring","https:\u002F\u002Fgithub.com\u002Fcarsimax\u002Fconstruct","https:\u002F\u002Fgithub.com\u002Fcarsimax","2026-10-30 09:50:40",{"source":8,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add carsimax\u002Fconstruct --skill agents-and-skills-authoring",{"tool":36,"command":37},"bunx","bunx skills add carsimax\u002Fconstruct --skill agents-and-skills-authoring",{"tool":39,"command":40},"pnpm","pnpm dlx skills add carsimax\u002Fconstruct --skill agents-and-skills-authoring","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fcarsimax\u002Fconstruct\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fagents-and-skills-authoring\u002FSKILL.md",true,[],[45,52,59,66,73,80],{"skillNo":46,"name":47,"nameZh":6,"description":48,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":49,"categories":50,"keywords":51,"publishedAt":19},"01M7F69PT3SMFA4AEETFGW0WFH","workflow","The ADD protocol for every implementation request, feature plan, resume, task sizing, and the closing Risk line.","https:\u002F\u002Fgithub.com\u002Fcarsimax\u002Fconstruct\u002Fblob\u002Fmain\u002Fplugins\u002Fconstruct\u002Fskills\u002Fworkflow\u002FSKILL.md",[],[],{"skillNo":53,"name":54,"nameZh":6,"description":55,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":56,"categories":57,"keywords":58,"publishedAt":19},"01M7F69J7R3HG57HMVY7QNQNX8","sources","Ground version-dependent code in official documentation. Use when you name a specific uncertainty about a library or framework API, a version-dependent pattern, or a deprecation. Skip when the repository already shows the pattern.","https:\u002F\u002Fgithub.com\u002Fcarsimax\u002Fconstruct\u002Fblob\u002Fmain\u002Fplugins\u002Fconstruct\u002Fskills\u002Fsources\u002FSKILL.md",[],[],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":19},"01M7F69CZ1YC0WHK90EV82CB1M","review","Mandatory outside review of high-risk work units, by lens agents the hooks feed and record, plus the dual blind review on request. Trigger: a unit whose Risk names an item is ready to publish, the publication gate denied a push, PR, or release, or the user asks for a review, a dual review, or an adversarial review.","https:\u002F\u002Fgithub.com\u002Fcarsimax\u002Fconstruct\u002Fblob\u002Fmain\u002Fplugins\u002Fconstruct\u002Fskills\u002Freview\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":19},"01M7F699XTX8F3PNZB0P86XH44","migration","Plan and verify risky migrations and removals. Use when changing a schema or stored data, removing or renaming a public API, flag, config format, or export, replacing a system, or when the task's Risk names item 1 or 3.","https:\u002F\u002Fgithub.com\u002Fcarsimax\u002Fconstruct\u002Fblob\u002Fmain\u002Fplugins\u002Fconstruct\u002Fskills\u002Fmigration\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":19},"01M7F6953SF387B7GX75V6MXE0","delivery","Plan commits, branches, and pull requests as reviewable units. Use when deciding commit contents, opening or preparing a PR, splitting large work into sub-branches of a feature branch, or preparing a release.","https:\u002F\u002Fgithub.com\u002Fcarsimax\u002Fconstruct\u002Fblob\u002Fmain\u002Fplugins\u002Fconstruct\u002Fskills\u002Fdelivery\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":19},"01M7F68YK43AQ6BM0KQPN0BVP4","delegation","When and how the ADD orchestrator delegates to construct agents: role mapping, allowed edit surfaces, parallel writers, verifier timing, handoff prompts, and relaying a subagent's blocking question or choice menu. Trigger: deciding to delegate, launching an agent, writing a handoff, or a subagent or tool returns a question, menu, or consent request.","https:\u002F\u002Fgithub.com\u002Fcarsimax\u002Fconstruct\u002Fblob\u002Fmain\u002Fplugins\u002Fconstruct\u002Fskills\u002Fdelegation\u002FSKILL.md",[],[]]