本文へ移動
cccskills
無料GitHub で公開

ase-spec-activate

Activate Specification Know-How: load the SpecBook format contract, the SpecBook schema configuration of the project, and the list of specification (SPEC) artifacts into the context, so the specification can be read, queried, explained, and edited ad-hoc in plain conversation. Use *automatically* whenever the user wants to work with the specification files -- asks what the "spec" or "specification" says, wants to "look up", "query", "check", or "explain" specification content, or wants a small ad-hoc change to it -- and no dedicated specification skill (`ase-spec-edit`, `ase-sync-import`, `ase-sync-reconcile`, `ase-sync-export`) is invoked, as those activate the know-how implicitly.

インストール方法を見る

含まれるファイル(2)

  • SKILL.md8.8 KB
  • help.md3.0 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

@${CLAUDE_SKILL_DIR}/../../meta/ase-control.md @${CLAUDE_SKILL_DIR}/../../meta/ase-skill.md @${CLAUDE_SKILL_DIR}/../../meta/ase-getopt.md

<purpose name="ase-spec-activate"> Activate Specification Know-How </purpose> <expand name="getopt" arg1="ase-spec-activate"> $ARGUMENTS </expand> <objective> *Activate* the know-how about the **SpecBook**-based specification -- the format contract, the schema configuration, and the artifact set -- so the specification can be worked with *ad-hoc* in plain conversation. </objective>

@${CLAUDE_SKILL_DIR}/../../meta/ase-format-meta.md @${CLAUDE_SKILL_DIR}/../../meta/ase-format-spec.md @${CLAUDE_SKILL_DIR}/../../meta/ase-tenets.md

Procedure

This skill is plan-less and read-mostly: it never composes or persists a task plan and MUST NOT call ase_task_save(...). It only loads know-how into the context and answers an optional query. It modifies SPEC artifacts only when the query explicitly asks for an ad-hoc change, and it never touches the artifact kinds CODE, DOCS, TASK, INFR, and OTHR.

  1. Initialize:

    Set <query><getopt-arguments/></query> (with any leading and trailing whitespace stripped). Do not output anything.

  2. Activate Format:

    The SpecBook format contract is already loaded into the context through the included ase-format-spec.md (and its nested ase-format-specbook.md). Internalize it now: the object kinds, the object ids and {{<id/>}} anchors, the [[xxx]] references, the , BECAUSE rationale split, the Complex/Concise/Grouped format variants, and the Created:/Modified: frontmatter block. Do not output anything.

  3. Activate Schema:

    1. Set <schema-files/> to the list of file paths constituting the SpecBook SCHEMA Model of the project (resolved as described in ase-format-spec.md), each annotated with standard for the bundled schema or custom for a project-specific one. Do not output anything.

    2. <if condition="the content of a file of <schema-files/> was *not* already read into the current context -- neither by an earlier run of this skill nor by a dedicated specification skill"> Read those files of the **SpecBook SCHEMA Model** in <schema-files/> via the `Read` tool to learn the allowed object kinds, properties, nestings, and value constraints. Do not output anything. </if> <else> Do *not* read the files of <schema-files/> again -- their content is already active in the context. Do not output anything. </else>
  4. Activate Artifacts:

    Resolve the SPEC artifacts by calling the ase_artifact_list(kind: [ "spec" ]) tool of the ase MCP server once and reading the returned artifacts array of { kind, files } objects. Set <artifact-files/> to the project-relative file list and <artifact-count/> to its length. Do not read the artifacts themselves in this step. Do not output anything.

  5. Report Activation:

    Only output the following <template/>:

    <template> <ase-tpl-boxed title="SPEC" subtitle="ACTIVATED">

    SCHEMA: <schema-files/> (one <file/> (standard|custom) entry per line, aligned) SPEC: <ase-spec-basedir/> (<artifact-count/> artifacts)

    </ase-tpl-boxed> </template>
  6. Answer Query:

    <if condition="<query/> is empty"> Silently *skip* this item. Do not output anything about the skipping. </if> <else> The <query/> argument *mainly* serves the *automatic* invocation by the agent harness, which passes the triggering user request *verbatim*, so activation and answer happen in *one* skill run.

    Serve <query/> ad-hoc under the Activated Behavior below: read the SPEC artifacts of <artifact-files/> which are related to <query/>, resolve their [[xxx]] references across artifacts, and answer the query grounded in the specification content, citing the artifact file and object id of every statement you rely on. If <query/> asks for a change, apply it as an ad-hoc modification according to the Activated Behavior. Set <answer/> to the resulting answer and only output the following <template/>:

    <template> <ase-tpl-bullet-normal/> **SPEC ANSWER**: <answer/> </template> </else>
  7. Finish:

    Finish the skill processing, but first give the closing hints by expanding the following (which, depending on the configured <ase-guidance-level/>, may each expand into nothing and hence emit no output at all):

    <ase-tpl-hint level="normal"> The specification know-how now stays active for this session: read, query, and ad-hoc edit the `SPEC` artifacts in plain conversation, or use `/ase-spec-edit` for a complete one-shot edit. </ase-tpl-hint> <ase-tpl-hint level="verbose"> Use `/ase-sync-import -t SPEC` to bring foreign sources into the specification, use `/ase-sync-reconcile -s SPEC` to propagate it into the other artifact kinds, and use `/ase-sync-export` to render it into HTML or PDF. </ase-tpl-hint>

Activated Behavior

Once this skill has run, the following rules stay in force for the remainder of the session whenever SPEC artifacts are read, queried, explained, or modified ad-hoc in plain conversation -- i.e. outside of any dedicated specification skill:

  • Resolution: You MUST resolve the SPEC artifacts via the ase_artifact_list(kind: [ "spec" ]) tool of the ase MCP server and never guess their file paths. Re-resolve them whenever artifacts might have been added or removed.

  • Interpretation: You MUST interpret the specification content strictly according to the SpecBook format contract and the SpecBook SCHEMA Model: object kinds, object ids and {{<id/>}} anchors, properties, [[xxx]] references, and , BECAUSE rationales. Before answering a query, resolve the [[xxx]] references across artifacts, and ground every answer in the specification content by citing artifact file and object id.

  • Modification: Every ad-hoc modification of a SPEC artifact MUST honor the GENERIC TENETS and the SPECIFYING TENETS of the ASE Tenets, and MUST keep the artifact conformant to the format contract and the schema: the Created:/Modified: frontmatter block, the heading levels, the format variants, the schema-allowed object kinds, nestings, and property keys, the object ids and anchors, the rationale split, and the references. Call the ase_timestamp(format: "yyyy-LL-dd HH:mm") tool of the ase MCP server once per change set and use its result to refresh the Modified: line of every changed artifact and for both the Created: and Modified: lines of every generated artifact.

  • Validation: After every ad-hoc modification, call the ase_specbook_lint() tool of the ase MCP server, fix the reported diagnostics in the affected SPEC artifacts for at most three rounds, and report any remaining diagnostics as <file/>:<line/>:<column/>: <message/> lines.

  • Restriction: An ad-hoc modification of the specification MUST stay restricted to the SPEC artifacts -- the artifact kinds CODE, DOCS, TASK, INFR, and OTHR are never touched as part of it. For a substantial or multi-artifact change, point the user to /ase-spec-edit instead of applying it ad-hoc.

  • Tandem: The dedicated specification skills ase-spec-edit, ase-sync-import, ase-sync-reconcile, and ase-sync-export include the format contract themselves and activate the know-how implicitly. Whenever one of them is invoked, its own procedure takes precedence over these rules, and this skill MUST NOT be invoked from within it. Conversely, a run of this skill never replaces such a dedicated skill when the user explicitly asks for it.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Review software architecture, including package cohesion and inter-package coupling

日本語の概要は準備中です。原文の説明を表示しています。

rse/ase592026年10月11日 更新

Discover additional, third-party components (libraries/frameworks) for the technology stack to provide needed functionality.

日本語の概要は準備中です。原文の説明を表示しています。

rse/ase592026年10月11日 更新

Analyze the source code for problems in either the logic and semantics and its related control flow, performance and efficiency, or security.

日本語の概要は準備中です。原文の説明を表示しています。

rse/ase592026年10月11日 更新

Craft Source Code: Use when user wants to "create", "add", or "craft" a new feature from scratch.

日本語の概要は準備中です。原文の説明を表示しています。

rse/ase592026年10月11日 更新

Edit Source Code: Use when the user wants to "edit" the code base in one shot from a query or a bare analyzer issue id like "P1", fusing crafting, refactoring, and resolving with optional grilling, verification, looping, and Git worktree isolation.

日本語の概要は準備中です。原文の説明を表示しています。

rse/ase592026年10月11日 更新

Explains code with WHAT, WHY, ANALOGY, DIAGRAM, CRUXES, and GOTCHAS. Use when you want to know how code works or when the user asks "how does this work?"

日本語の概要は準備中です。原文の説明を表示しています。

rse/ase592026年10月11日 更新

rse のスキルをすべて見る

このスキルの問題を報告する