Skip to content

rse/ase / plugin/skills/ase-sync-reconcile

ase-sync-reconcile

Verified

Reconcile one set of artifact kinds (the target) to reflect the current state of another set of artifact kinds (the source), while optionally honoring a filtering hint. Use when the user wants to "reconcile", "sync", "align", or "update" artifacts like SPEC, CODE, DOCS, TASK, INFR, or OTHR against each other.

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

$ARGUMENTS

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

Procedure

    1. The recognized artifact kinds are the six tokens TASK, SPEC, CODE, DOCS, INFR, and OTHR. Parse as the comma-separated kind list and as the comma-separated kind list. Upper-case and trim every parsed kind token. Do not output anything.

    2. Only output the following and then immediately STOP processing the entire current skill:

    3. Set to the six recognized kinds TASK,SPEC,CODE,DOCS,INFR,OTHR minus all kinds present in . Do not output anything.

    4. If any token in or is not one of the six recognized kinds, only output the following (with set to the first offending token) and then immediately STOP processing the entire current skill:

    5. Remove from any kind that is also present in (a kind is never its own source).

      Only output the following and then immediately STOP processing the entire current skill:

    6. Treat as a comma-separated list of operation tokens. The getopt parser validates only the first token, so you MUST validate each remaining token yourself against the allowed set all, add, update, remove. If any token is not in this set, bind to that offending token, then only output the following and then immediately STOP processing the entire current skill:

      Set to add,update,remove if all is in , or to the given tokens otherwise. Do not output anything.

    7. Report the resolved target, source, and operations with the following :

    1. Do not output anything in this STEP 2.

    2. For all kinds in the union of and , call the ase_artifact_list(kind: [ ... ]) tool of the ase MCP server once, passing the lower-cased kind tokens, and read the returned artifacts array of { kind, files } objects to obtain the project-relative file list per kind.

    3. Honor the filtering to reduce the source and/or target artifacts and/or the aspects of those artifacts you should take into account.

    4. Read all (optionally filtered) source and target artifacts previously resolved and build a precise understanding of the current state they represent.

    1. Internalize and honor the artifact-format conventions:

      • the artifact-set meta information (ase-format-meta.md),
      • the SPEC format, i.e. the SpecBook models and formats plus the SpecBook schema configuration of the project (ase-format-spec.md),
      • the TASK format (ase-format-task.md).

      Whenever a target artifact belongs to one of these kinds, the update MUST keep it conformant to the corresponding format (headings, structure, identifiers, references, and the Modified: timestamp rule). The kinds CODE, DOCS, INFR, and OTHR have no dedicated format contract and are treated as free-form.

    2. You MUST internalize and strictly honor the GENERIC TENETS, the RECONCILIATION TENETS, the REFACTORING TENETS, and the CRAFTING TENETS of the ASE Tenets when updating in the following. Do not output anything.

    3. Call the ase_timestamp(format: "yyyy-LL-dd HH:mm") tool of the ase MCP server once to find out the current time and store it in .

    4. Bidirectionally update the and artifacts so that they faithfully reflect the current state of each other.

      Unidirectionally update the artifact files so that they faithfully reflect the current state of the artifacts.

      Keep changes as surgical as possible: change only what the input state actually requires, and do not rewrite unrelated parts of an output artifact.

      Restrict the update to the operations in , on the level of both the content inside an output artifact and whole output artifact files: add creates output content and files the input warrants but the output lacks, update changes existing output content the input contradicts, and remove deletes output content and files the input no longer supports. Whatever an excluded operation would have changed stays silently untouched.

      For each formatted output artifact kind, strictly honor its format contract.

      Whenever an output artifact is changed and contains a Modified: <timestamp-modified-old/> line, replace this with Modified: <timestamp-modified/>.

      Do not modify any output artifact: you MUST NOT call Edit, Write, NotebookEdit, ase_task_save, or any other filesystem-modifying tool or shell command. Instead, store the complete change set in unified diff format in , diffing every touched file against its current on-disk content and every touched TASK output against its plan text loaded via the ase_task_load MCP tool, so all context lines match exactly and the diff would apply cleanly.

      Apply the update directly to the output artifacts via the Write/Edit tools. For a TASK output, apply it via the ase_task_save MCP tool instead -- NEVER write a task plan file via Write/Edit or by executing a shell command.

    5. Validate the specification by calling the ase_specbook_lint() tool of the ase MCP server and reading its returned diagnostics array of { file, line, column, severity, message } objects. If it is not empty, fix the reported problems in the affected SPEC artifacts via the Write/Edit tools and call the tool again -- for at most three rounds in total. Do not output anything, unless diagnostics remain after the last round, in which case output the following , listing one bullet line per remaining diagnostic:

      • <file/>:<line/>:<column/>: [...]
    6. Report the intended updates with the following . Set to a run of backtick characters one longer than the longest backtick run occurring anywhere inside , but to at least three, so a diff which itself carries fenced code blocks cannot terminate the block prematurely:

      diff

      Only output the following :

      Report the performed updates with the following , listing one bullet line per changed output file (with its project-relative path and an ultra-brief description of what was reconciled):

      • <file/>: [...]

      Only output the following :

    7. Finally, give the closing hints by expanding the following (which, depending on the configured , may each expand into nothing and hence emit no output at all):

rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization.Stars: 468kUpdated Claude CodeCodexVerifiedInstall
academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.Stars: 197kUpdated Claude CodeCodexVerifiedInstall
claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file).Stars: 197kUpdated Claude CodeCodexUnverifiedInstall
doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.Stars: 197kUpdated Claude CodeCodexVerifiedInstall