dan:unify — for Claude Code dan:unify, community, for Claude Code, ide skills, {phase}-{plan}-SUMMARY.md, If status is, IN_PROGRESS, APPROVED, dan:apply, <objective>

v1.0.0

关于此技能

适用场景: Ideal for AI agents that need reads the target plan.md and its task commits. 本地化技能摘要: My personal PAUL / CARL # dan:unify Close the execution loop for a completed plan. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

功能特性

Reads the target PLAN.md and its task commits
Gathers deviations, decisions, and issues from execution
Produces {phase}-{plan}-SUMMARY.md with:
Frontmatter (dependency graph, tech-stack, key-files, decisions, metrics)
Accomplishments, task commits, files created/modified

# 核心主题

RavBogard RavBogard
[0]
[0]
更新于: 3/29/2026

Killer-Skills Review

Decision support comes first. Repository text comes second.

Reference-Only Page Review Score: 8/11

This page remains useful for operators, but Killer-Skills treats it as reference material instead of a primary organic landing page.

Original recommendation layer Concrete use-case guidance Explicit limitations and caution
Review Score
8/11
Quality Score
46
Canonical Locale
en
Detected Body Locale
en

适用场景: Ideal for AI agents that need reads the target plan.md and its task commits. 本地化技能摘要: My personal PAUL / CARL # dan:unify Close the execution loop for a completed plan. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

核心价值

推荐说明: dan:unify helps agents reads the target plan.md and its task commits. My personal PAUL / CARL # dan:unify Close the execution loop for a completed plan. This AI agent skill supports Claude Code, Cursor, and

适用 Agent 类型

适用场景: Ideal for AI agents that need reads the target plan.md and its task commits.

赋予的主要能力 · dan:unify

适用任务: Applying Reads the target PLAN.md and its task commits
适用任务: Applying Gathers deviations, decisions, and issues from execution
适用任务: Applying Produces {phase}-{plan}-SUMMARY.md with:

! 使用限制与门槛

  • 限制说明: Requires repository-specific context from the skill documentation
  • 限制说明: Works best when the underlying tools and dependencies are already configured

Why this page is reference-only

  • - Current locale does not satisfy the locale-governance contract.
  • - The underlying skill quality score is below the review floor.

Source Boundary

The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.

评审后的下一步

先决定动作,再继续看上游仓库材料

Killer-Skills 的主价值不应该停在“帮你打开仓库说明”,而是先帮你判断这项技能是否值得安装、是否应该回到可信集合复核,以及是否已经进入工作流落地阶段。

实验室 Demo

Browser Sandbox Environment

⚡️ Ready to unleash?

Experience this Agent in a zero-setup browser environment powered by WebContainers. No installation required.

Boot Container Sandbox

常见问题与安装步骤

以下问题与步骤与页面结构化数据保持一致,便于搜索引擎理解页面内容。

? FAQ

dan:unify 是什么?

适用场景: Ideal for AI agents that need reads the target plan.md and its task commits. 本地化技能摘要: My personal PAUL / CARL # dan:unify Close the execution loop for a completed plan. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

如何安装 dan:unify?

运行命令:npx killer-skills add RavBogard/DAN。支持 Cursor、Windsurf、VS Code、Claude Code 等 19+ IDE/Agent。

dan:unify 适用于哪些场景?

典型场景包括:适用任务: Applying Reads the target PLAN.md and its task commits、适用任务: Applying Gathers deviations, decisions, and issues from execution、适用任务: Applying Produces {phase}-{plan}-SUMMARY.md with:。

dan:unify 支持哪些 IDE 或 Agent?

该技能兼容 Cursor, Windsurf, VS Code, Trae, Claude Code, OpenClaw, Aider, Codex, OpenCode, Goose, Cline, Roo Code, Kiro, Augment Code, Continue, GitHub Copilot, Sourcegraph Cody, and Amazon Q Developer。可使用 Killer-Skills CLI 一条命令通用安装。

dan:unify 有哪些限制?

限制说明: Requires repository-specific context from the skill documentation;限制说明: Works best when the underlying tools and dependencies are already configured。

安装步骤

  1. 1. 打开终端

    在你的项目目录中打开终端或命令行。

  2. 2. 执行安装命令

    运行:npx killer-skills add RavBogard/DAN。CLI 会自动识别 IDE 或 AI Agent 并完成配置。

  3. 3. 开始使用技能

    dan:unify 已启用,可立即在当前项目中调用。

! 参考页模式

此页面仍可作为安装与查阅参考,但 Killer-Skills 不再把它视为主要可索引落地页。请优先阅读上方评审结论,再决定是否继续查看上游仓库说明。

Upstream Repository Material

The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.

Upstream Source

dan:unify

My personal PAUL / CARL # dan:unify Close the execution loop for a completed plan. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

SKILL.md
Readonly
Upstream Repository Material
The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.
Supporting Evidence

dan:unify

Close the execution loop for a completed plan. Reads task commits, deviations, and verification results to produce a SUMMARY.md file. Updates STATE.md and ROADMAP.md.

What It Does

  1. Reads the target PLAN.md and its task commits
  2. Gathers deviations, decisions, and issues from execution
  3. Produces {phase}-{plan}-SUMMARY.md with:
    • Frontmatter (dependency graph, tech-stack, key-files, decisions, metrics)
    • Accomplishments, task commits, files created/modified
    • Deviations from plan, issues encountered
  4. Updates STATE.md (advance plan counter, add decisions, record metrics)
  5. Updates ROADMAP.md (plan progress row)

Execution Mode

Runs in-session (no agent spawning). Reads git log and plan files to produce summary.

CLI Tools

State and roadmap operations use the DAN CLI:

bash
1node "$HOME/.claude/dan/bin/dan-tools.cjs" --cwd $PROJECT_DIR state set "Plan" "3 of 4 in current phase" 2node "$HOME/.claude/dan/bin/dan-tools.cjs" --cwd $PROJECT_DIR state get "Status"

<execution_flow>

<step name="identify_plan"> Determine which plan to unify (close the loop on).
  1. If a plan path argument was provided by the user, use it directly:

    PLAN_FILE=$1   # e.g., ".planning/phases/02-core-loop/02-01-PLAN.md"
    
  2. Otherwise, read STATE.md for current phase and plan position:

    bash
    1node "$HOME/.claude/dan/bin/dan-tools.cjs" --cwd "$PROJECT_DIR" state get "Phase" 2node "$HOME/.claude/dan/bin/dan-tools.cjs" --cwd "$PROJECT_DIR" state get "Plan"

    Determine the phase directory and scan for unclosed plans.

  3. Find the most recent PLAN.md without a matching SUMMARY.md:

    bash
    1PHASE_PATH="$PROJECT_DIR/.planning/phases/${PADDED_PHASE}-${PHASE_NAME}" 2for PLAN in $(ls "$PHASE_PATH"/*-PLAN.md 2>/dev/null); do 3 PREFIX=$(basename "$PLAN" | sed 's/-PLAN.md//') 4 SUMMARY="$PHASE_PATH/${PREFIX}-SUMMARY.md" 5 [ ! -f "$SUMMARY" ] && PLAN_FILE="$PLAN" && break 6done
  4. Parse the plan frontmatter and verify status:

    bash
    1node "$HOME/.claude/dan/bin/dan-tools.cjs" --cwd "$PROJECT_DIR" frontmatter parse "$PLAN_FILE"
    • If status is COMPLETED: proceed (summary is the missing piece)
    • If status is IN_PROGRESS: warn "Plan is still IN_PROGRESS. Producing summary anyway -- verify all tasks are done."
    • If status is DRAFT or APPROVED: error "Plan has not been executed yet. Run /dan:apply first."
  5. If no unclosed plans found:

    All plans in phase ${PADDED_PHASE} have summaries. Nothing to unify.
    
</step> <step name="read_plan_and_commits"> Gather all execution evidence for the plan.
  1. Read the PLAN.md file and extract:

    • Objective (from <objective> block)
    • Tasks (from <task> elements): name, type, files, done criteria
    • Success criteria (from <success_criteria> block)
    • Requirements covered (from frontmatter requirements)
    • Files modified (from frontmatter files_modified)
  2. Read git log for commits matching this plan's identifier:

    bash
    1git log --oneline --grep="(${PADDED_PHASE}-${PADDED_PLAN})" --all

    This captures all task commits (e.g., feat(02-01): ..., test(02-01): ...).

  3. For each task in the plan, determine its status:

    • Completed: A commit exists with the task's description or file changes
    • Skipped: No matching commit and task was explicitly deferred
    • Modified: Commit exists but files or scope differ from plan
  4. Check the current state of planned files:

    bash
    1for FILE in ${FILES_MODIFIED}; do 2 [ -f "$PROJECT_DIR/$FILE" ] && echo "EXISTS: $FILE" || echo "MISSING: $FILE" 3done
  5. Count actual files created/modified from the git diff:

    bash
    1git diff --name-only $(git log --oneline --grep="(${PADDED_PHASE}-${PADDED_PLAN})" --format="%H" | tail -1)^..$(git log --oneline --grep="(${PADDED_PHASE}-${PADDED_PLAN})" --format="%H" | head -1)
</step> <step name="gather_qualification_results"> Collect qualification and deviation data from the execution.
  1. Check STATE.md decisions section for any qualification results:

    bash
    1node "$HOME/.claude/dan/bin/dan-tools.cjs" --cwd "$PROJECT_DIR" state get "Decisions"

    Look for entries mentioning this plan's identifier or PASS_WITH_CONCERNS.

  2. Check for deviation patterns in commit messages:

    bash
    1git log --oneline --grep="(${PADDED_PHASE}-${PADDED_PLAN})" --all | grep -iE "fix|deviat|missing|bug"
  3. Collect any PASS_WITH_CONCERNS notes -- these become items in the "Deviations" or "Deferred" sections.

  4. Note any auto-fixed bugs, added missing functionality, or blocking issues resolved during execution. These map to deviation rules:

    • Rule 1 (Bug fixes)
    • Rule 2 (Missing critical functionality)
    • Rule 3 (Blocking issues) </step>
<step name="produce_summary"> Create the SUMMARY.md file using the summary template structure.
  1. Determine the summary file path:

    bash
    1SUMMARY_FILE="$PHASE_PATH/${PADDED_PHASE}-${PADDED_PLAN}-SUMMARY.md"
  2. Build the summary content following bin/templates/summary.md structure:

    Frontmatter:

    yaml
    1--- 2phase: {PADDED_PHASE} 3plan: {PADDED_PLAN} 4completed: {YYYY-MM-DD} 5---

    Title: # Phase {PHASE} Plan {PLAN}: [Name] Summary

    One-liner: Substantive description of what was built (not "X implemented" but "JWT auth with refresh rotation using jose library").

    Sections:

    • Objective: Copied from the plan's <objective> block
    • What Was Built: List actual deliverables with file paths from git evidence
    • Plan vs Actual: Table comparing:
      AspectPlannedActual
      Tasks{planned_count}{actual_count}
      Files{planned_files}{actual_files}
      Duration{estimated}{actual_duration}
    • Decisions Made: Any decisions logged during execution (from STATE.md or commit messages)
    • Deviations from Plan: Differences between planned and actual work. If none: "None - plan executed exactly as written."
    • Deferred Items: Anything planned but not completed. If none: "None."
  3. Write the summary file: Use the Write tool to create $SUMMARY_FILE with the assembled content.

    </step>
<step name="update_state_and_roadmap"> Transition the plan to COMPLETED and update all tracking files.
  1. Transition the plan status to COMPLETED:

    bash
    1node "$HOME/.claude/dan/bin/dan-tools.cjs" --cwd "$PROJECT_DIR" frontmatter set "$PLAN_FILE" status COMPLETED

    The lifecycle module validates the transition (IN_PROGRESS -> COMPLETED or COMPLETED -> COMPLETED is idempotent).

  2. Update STATE.md -- advance the plan counter:

    bash
    1node "$HOME/.claude/dan/bin/dan-tools.cjs" --cwd "$PROJECT_DIR" state set "Plan" "${NEXT_PLAN_NUM} of ${TOTAL_PLANS} in current phase" 2node "$HOME/.claude/dan/bin/dan-tools.cjs" --cwd "$PROJECT_DIR" state set "Last activity" "$(date +%Y-%m-%d) -- Completed plan ${PADDED_PHASE}-${PADDED_PLAN}"
  3. Add any new decisions to STATE.md:

    bash
    1node "$HOME/.claude/dan/bin/dan-tools.cjs" --cwd "$PROJECT_DIR" state get "Decisions"

    Append new decisions from the execution if not already present.

  4. Update ROADMAP.md -- mark the plan as complete: Find the plan's entry in ROADMAP.md and change [ ] to [x]:

    bash
    1# Read ROADMAP.md, find the line for this plan, mark complete 2# Use Edit tool to change "- [ ] Plan XX-NN" to "- [x] Plan XX-NN"
  5. Commit all changes:

    bash
    1node "$HOME/.claude/dan/bin/dan-tools.cjs" --cwd "$PROJECT_DIR" commit "unify(${PADDED_PHASE}-${PADDED_PLAN}): close loop with summary" --files "$SUMMARY_FILE" "$PROJECT_DIR/.planning/STATE.md" "$PROJECT_DIR/.planning/ROADMAP.md" "$PLAN_FILE"
</step> <step name="offer_next"> Suggest the appropriate next action based on phase progress.
  1. Check if more plans exist in the current phase:

    bash
    1TOTAL_PLANS=$(ls "$PHASE_PATH"/*-PLAN.md 2>/dev/null | wc -l) 2COMPLETED_PLANS=$(ls "$PHASE_PATH"/*-SUMMARY.md 2>/dev/null | wc -l)
  2. If uncompleted plans remain:

    Loop closed for plan ${PADDED_PHASE}-${PADDED_PLAN}.
    Next: Run `/dan:apply` on plan ${PADDED_PHASE}-${NEXT_PADDED_PLAN} to continue.
    
  3. If all plans in the phase are complete:

    Phase ${PADDED_PHASE} complete -- all plans have summaries.
    Next: Run `/dan:verify` to validate phase deliverables, or `/dan:plan` for the next phase.
    
  4. Always confirm closure:

    The loop is now closed for plan ${PADDED_PHASE}-${PADDED_PLAN}.
    
</step>

</execution_flow>

相关技能

寻找 dan:unify 的替代方案 (Alternative) 或可搭配使用的同类 community Skill?探索以下相关开源技能。

查看全部

openclaw-release-maintainer

Logo of openclaw
openclaw

Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

333.8k
0
AI

widget-generator

Logo of f
f

为prompts.chat的信息反馈系统生成可定制的插件小部件

149.6k
0
AI

flags

Logo of vercel
vercel

React 框架

138.4k
0
浏览器

pr-review

Logo of pytorch
pytorch

Python中具有强大GPU加速的张量和动态神经网络

98.6k
0
开发者工具