document-planner
Interactive planning at any scope — whole thesis, chapter, section, or subsection. Invoke with a prompt specifying the working level. Collaboratively builds writing plans from cold start through to statement-level outlines with verified Zotero citations.
What this skill does
# Document Planner ## Overview This skill collaboratively builds writing plans at any scope — from whole-thesis chapter maps down to statement-level paragraph outlines with verified citations. It is the single planning skill for all document types and hierarchy levels. The process is **fundamentally collaborative**. The author is a subject-matter expert who knows what points are needed to communicate clearly. You are a writing expert who structures documents logically, divides responsibilities into sections, and maintains narrative coherence across the wider document. Neither party works autonomously — every structural decision is discussed. **CRITICAL — Zotero access policy**: NEVER call `mcp__deep-zotero__*` tools directly. All Zotero library access MUST go through the `zotero-research` agent, spawned via the Task tool. Only the `zotero-research` agent is permitted to call the MCP tools. ## When to Use This Skill - Planning a new thesis from scratch (whole-thesis scope) - Planning individual chapters from a thesis-level plan - Building paragraph-level or statement-level plans for sections - Revising the structure of an already-planned section - Planning a standalone paper or report ## Chapter Types Each chapter type has a distinct structure and writing approach. Use these when proposing chapter plans. ### Background Chapter Literature-review style treatment. Provides context and foundation. - **Structure**: Introduction → Major themes (organised thematically, NOT study-by-study) → Summary with gaps - **Writing approach**: Synthesize across sources. Every claim needs citation(s). Identify gaps and controversies. Connect to thesis objectives. - **Citation density**: High (1-3 citations per paragraph) - **Figures**: Conceptual diagrams, timelines, comparison tables ### Meat Chapter (Research Chapter) Paper-like IMRaD structure presenting original research. Each should be publishable as standalone. - **Structure**: Introduction → Brief background (reference full background chapter) → Methods → Results → Discussion → Conclusion - **Writing approach**: Self-contained but connected to thesis narrative. Methods fully reproducible. Results before interpretation. - **Citation density**: Medium in intro/discussion, low in methods/results (own work) - **Figures**: Data plots, block diagrams, result visualizations ### Conclusions Chapter Synthesis of thesis findings and contributions. - **Structure**: Summary of work → Key findings → Contributions → Limitations - **Writing approach**: No new material. Reference thesis chapters, not external sources. Specific about contributions. Honest about limitations. - **Citation density**: None (reference thesis chapters only) ### Future Work Chapter Research directions and open questions. - **Structure**: Immediate extensions → Longer-term directions → Open questions - **Writing approach**: Concrete and actionable. Connect to thesis limitations. - **Citation density**: Low ## Inputs 1. **Target scope**: Whole thesis, chapter, section, or subsection 2. **Document directory**: The directory containing the `.tex` file and (optionally) a `plan.md` ## Document Hierarchy Three documents form an authority hierarchy for content. Higher-level documents set narrative and structure; lower-level documents add detail: 1. **`.tex` file** (most detailed, authoritative for existing prose) 2. **Directory-level `plan.md`** (paragraph/statement-level plan, created/maintained by this skill) 3. **Parent-level `plan.md`** (chapter-level or thesis-level plan, sets narrative goals) If there is a `plan.md` one directory above the document directory, treat it as the hierarchical master plan for this document. **Parent plan detail level:** The thesis-level plan should contain section purposes, high-level bullet points, and narrative goals. It does not contain references — those are resolved at the directory-level plan during Phase 3. The document-planner may update the parent plan to reflect structural changes (section reordering, new/removed sections, cross-chapter scope changes) but must NOT inflate it with paragraph-level detail, figure descriptions, or expanded sub-points. Those belong in the directory-level plan only. ### Startup: Read and Reconcile 1. **Read the `.tex` file first.** Existing content is authoritative — it must not be removed without explicit user discussion. The user edits this file between sessions. 2. **Read the directory-level `plan.md`** if it exists. This is the detailed plan from previous sessions. The user may have modified it between sessions. 3. **Read the parent-level `plan.md`** (e.g., `thesis/plan.md`). This contains the high-level narrative and key points. **If the directory-level `plan.md` does not exist**, create one by copying in the parent plan's content for this document as a starting point. ### Reconciliation Rules Compare the three documents and resolve discrepancies before proceeding: - **Content in a higher-level doc missing from a lower-level doc**: Ask the user whether to (a) remove it from the higher-level doc, or (b) add it to the lower-level doc(s) in this or a future round. - **Content in a lower-level doc that diverges from or modifies the narrative/key points in a higher-level doc**: Ask the user whether to (a) update the higher-level doc to match the new narrative, or (b) modify the lower-level doc to fit the existing narrative. This hierarchy applies outside thesis structures too. Any document directory should have a `plan.md` and a `.tex` file; create them if missing. ### Structural Mismatches When section headings, ordering, or structure differ between documents (e.g., `.tex` has different subsection titles than `plan.md`), **ask the user for guidance** — do not silently pick one. Present the mismatch and ask which version to follow, then update the other document(s) to match. ### Understanding the Parent Plan The parent-level plan is a **thin outline** — a best attempt at a high-level overview of points that will meet narrative goals. It was typically built collaboratively with the author, so it **lacks specific detail**. Treat it as a structural starting point, not a content-complete source. ### Cold Start (No Higher-Level Plan) When there is no parent-level `plan.md`, build the thin starting plan collaboratively: 1. Ask the author what this document needs to achieve and who the audience is 2. Propose a section structure with narrative bullet points based on the author's input 3. Iterate with the author until the high-level structure is agreed 4. Write this as the directory-level `plan.md` and proceed to Phase 2 For whole-thesis cold starts, first establish the overall narrative (what is the thesis about, what is the contribution, what story does it tell), then map chapters (background, meat, conclusions) with their connections before planning individual chapters. ## Workflow ### Phase 1: Read and Reconcile Follow the Startup procedure above. Report what was found in each document and any discrepancies requiring user input. ### Phase 2: Narrative Structure and Paragraph Flow **This is the core collaborative phase.** Before touching references or detailed points, establish the narrative arc and paragraph flow for each section. #### Top-Down Planning Across Hierarchy Levels Documents have multiple levels of hierarchy (thesis → chapter → section → subsection → paragraph). **Plan top-down**: complete the narrative arc and prerequisite check at the highest level before descending to plan individual children at the next level. Then work through children sequentially, planning each one at its own level before descending further. **Procedure:** 1. **Plan the highest level first.** For a thesis, this means chapter ordering. For a chapter, section-level narrative arc. 2. **Get author agreement** on this level before proceeding. 3. **Descend one level.** Work through children **sequentially** — finish one child's plan before starting the next. 4. **Repeat**
Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.