wtf-p 0.3.0 → 0.5.0

package/vendors/claude/commands/wtfp/polish-prose.md

@@ -1,6 +1,6 @@
 ---
 name: wtfp:polish-prose
-description: De-robotize and refine prose for clarity, flow, and academic voice
+description: Improve clarity, flow, and academic voice in written prose
 argument-hint: "[section-number or file]"
 allowed-tools:
   - Read
@@ -9,66 +9,47 @@ allowed-tools:
   - Edit
   - Glob
   - Grep
+  - Task
   - AskUserQuestion
 ---
 
+<execution_context>
+@~/.claude/write-the-f-paper/references/principles.md
+</execution_context>
+
 <objective>
 Polish prose to eliminate robotic AI-sounding text and improve academic voice.
 
-Purpose: Transform draft prose into publication-ready writing that sounds human, flows naturally, and maintains academic rigor.
-Output: Refined section text with improved clarity, varied sentence structure, and natural transitions.
+**Orchestrator role:** Identify target files, gather voice preference, resolve model profile, spawn prose-polisher agent, present results.
 </objective>
 
-<execution_context>
-@~/.claude/write-the-f-paper/references/prose-quality.md
-@~/.claude/write-the-f-paper/references/academic-voice.md
-</execution_context>
-
 <context>
 Target: $ARGUMENTS (section number, file path, or "all" for full manuscript)
 
 **Load project state:**
 @.planning/STATE.md
-
-**Load style context if exists:**
-@.planning/sources/style-guide.md
 </context>
 
 <process>
 
-<step name="identify_target">
-1. Parse $ARGUMENTS to determine scope:
-   - Section number → find section draft in .planning/sections/XX-name/
-   - File path → use directly
-   - "all" or empty → process all draft sections
-2. Read target content
-3. Identify prose issues
-</step>
-
-<step name="analyze_prose">
-**Identify common AI-writing patterns to fix:**
-
-| Pattern | Fix |
-|---------|-----|
-| "It is important to note that..." | Delete or integrate naturally |
-| "This study aims to..." | Active voice: "We investigate..." |
-| "In conclusion, it can be said..." | Direct statement |
-| Excessive hedging | Confident assertions where warranted |
-| Repetitive sentence starts | Vary structure |
-| Overuse of "Furthermore", "Moreover" | Natural transitions |
-| Passive voice overload | Active where appropriate |
-| Nominalization | Verbs over noun forms |
-| Word padding | Tighten |
-
-**Check for:**
-- Sentence length variety (mix short punchy with longer complex)
-- Paragraph flow and transitions
-- Topic sentences that preview content
-- Concluding sentences that land
-- Jargon density appropriate for venue
-</step>
-
-<step name="gather_preferences">
+## 1. Validate Environment and Resolve Model Profile
+
+```bash
+ls paper/*.md 2>/dev/null || echo "No content in paper/ directory"
+```
+
+```bash
+MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
+```
+
+| Agent | quality | balanced | budget |
+|-------|---------|----------|--------|
+| prose-polisher | opus | sonnet | haiku |
+
+## 2. Identify Target and Gather Preferences
+
+Parse $ARGUMENTS to find target file(s).
+
 Use AskUserQuestion:
 - header: "Voice"
 - question: "What tone should dominate this section?"
@@ -77,101 +58,68 @@ Use AskUserQuestion:
   - "Measured" — Careful hedging, academic caution
   - "Accessible" — Clear to non-specialists
   - "Technical" — Dense, specialist terminology OK
-</step>
-
-<step name="polish">
-**Apply transformations:**
 
-1. **Sentence-level:**
-   - Eliminate filler phrases
-   - Convert passive to active where natural
-   - Vary sentence openings
-   - Tighten word count (cut 10-15%)
+## 3. Read Context Files
 
-2. **Paragraph-level:**
-   - Ensure clear topic sentences
-   - Check logical flow between sentences
-   - Verify transitions connect ideas
-
-3. **Section-level:**
-   - Opening hook engages
-   - Closing paragraph lands the point
-   - Section advances the argument
+```bash
+TARGET_CONTENT=$(cat paper/[target].md 2>/dev/null)
+STYLE_CONTENT=$(cat .planning/sources/style-guide.md 2>/dev/null)
 
-**Preserve:**
-- Technical accuracy
-- Citation placement
-- Author's core arguments
-- Discipline-specific terminology
-</step>
+SECTION_DIR=$(ls -d .planning/sections/${SECTION}-* 2>/dev/null | head -1)
+CONTEXT_CONTENT=$(cat "${SECTION_DIR}"/*-CONTEXT.md 2>/dev/null)
+```
 
-<step name="present_changes">
-Show before/after for significant changes:
+## 4. Spawn wtfp-prose-polisher Agent
 
 ```
-Before: "It is important to note that the results demonstrate..."
-After:  "The results demonstrate..."
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+ WTF-P ► POLISHING PROSE
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+```
 
-Before: "This methodology was utilized in order to..."
-After:  "We used this methodology to..."
+```
+Task(
+  prompt="First, read ~/.claude/agents/wtfp/prose-polisher.md for your role and instructions.\n\n" + filled_polish_prompt,
+  subagent_type="general-purpose",
+  model="{polisher_model}",
+  description="Polish prose"
+)
 ```
 
-Use AskUserQuestion:
-- header: "Review"
-- question: "Apply these prose refinements?"
-- options:
-  - "Apply all" — Looks good, update the file
-  - "Show more" — I want to see more examples
-  - "Selective" — Let me choose which changes
-</step>
+Polish prompt includes: `<target_content>` with section text, `<voice_preference>` with user selection, `<style_guide>` with style context, `<user_decisions>` with CONTEXT_CONTENT.
 
-<step name="apply">
-Apply approved changes to the target file(s).
-</step>
+## 5. Handle Polisher Return
 
-<step name="commit">
-```bash
-git add [modified files]
-git commit -m "$(cat <<'EOF'
-style(XX): polish prose for clarity and flow
-
-- Eliminated AI-sounding phrases
-- Varied sentence structure
-- Tightened word count
-- Improved transitions
-EOF
-)"
-```
-</step>
+**`## POLISH COMPLETE`:** Present before/after summary.
 
-<step name="done">
-```
-Prose polished:
+**`## POLISH BLOCKED`:** Present blocker, ask user how to proceed.
 
-- File(s): [list]
-- Word count: [before] → [after] ([X]% reduction)
-- Patterns fixed: [count]
+## 6. Present Results
 
----
+</process>
+
+<offer_next>
+
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+ WTF-P ► PROSE POLISHED ✓
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+───────────────────────────────────────────
 
 ## Next Up
 
-Review the changes, then continue:
-- `/wtfp:review-section [N]` — Get reviewer feedback
+- `/wtfp:review-section {N}` — Get reviewer feedback
 - `/wtfp:check-refs` — Verify citations
 - `/wtfp:export-latex` — Generate LaTeX
 
----
-```
-</step>
+───────────────────────────────────────────
 
-</process>
+</offer_next>
 
 <success_criteria>
-- AI-sounding patterns eliminated
-- Sentence variety improved
-- Transitions flow naturally
-- Word count tightened without losing meaning
-- Academic voice maintained
-- Changes committed to git
+- [ ] Target file(s) identified
+- [ ] Voice preference gathered
+- [ ] Prose-polisher spawned with full context
+- [ ] Changes applied to paper/ files
+- [ ] User knows next steps
 </success_criteria>

package/vendors/claude/commands/wtfp/progress.md

@@ -1,325 +1,153 @@
 ---
 name: wtfp:progress
-description: Check writing progress, show context, and route to next action
+description: Show writing progress and suggest next step
 allowed-tools:
   - Read
   - Bash
-  - Grep
   - Glob
-  - SlashCommand
+  - Grep
+  - Task
+  - AskUserQuestion
 ---
 
 <objective>
-Check writing progress, summarize recent work and what's ahead, then intelligently route to the next action - either executing an existing plan or creating the next one.
+Check writing progress, summarize recent work, and intelligently route to the next action.
 
-Provides situational awareness before continuing writing.
+**Orchestrator role:** Load state files, calculate progress metrics, present rich status report, determine and offer the next logical command.
 </objective>
 
+<context>
+No arguments. Reads project state from .planning/.
+</context>
 
 <process>
 
-<step name="verify">
-**Verify planning structure exists:**
-
-If no `.planning/` directory:
-
-```
-No planning structure found.
-
-Run /wtfp:new-paper to start a new paper.
-```
-
-Exit.
-
-If missing STATE.md or ROADMAP.md: inform what's missing, suggest running `/wtfp:new-paper`.
-</step>
-
-<step name="load">
-**Load full project context:**
-
-- Read `.planning/STATE.md` for living memory (position, word counts, decisions)
-- Read `.planning/ROADMAP.md` for section structure and objectives
-- Read `.planning/PROJECT.md` for paper vision (Core Argument, Requirements)
-</step>
-
-<step name="recent">
-**Gather recent work context:**
-
-- Find the 2-3 most recent SUMMARY.md files
-- Extract from each: what was written, key decisions, any issues logged
-- This shows "what we've been working on"
-</step>
-
-<step name="position">
-**Parse current position:**
-
-- From STATE.md: current section, plan number, status, word count
-- Calculate: total plans, completed plans, remaining plans
-- Calculate: total words written vs target word count
-- Note any gaps, concerns, or deferred issues
-- Check for CONTEXT.md: For sections without PLAN.md files, check if `{section}-CONTEXT.md` exists
-</step>
-
-<step name="report">
-**Present rich status report:**
-
-```
-# [Paper Title]
-
-**Progress:** [████████░░] 8/10 sections drafted
-**Words:** [current] / [target] ([percentage]%)
-
-## Recent Work
-- [Section X, Plan Y]: [what was written - 1 line]
-- [Section X, Plan Z]: [what was written - 1 line]
-
-## Current Position
-Section [N] of [total]: [section-name]
-Plan [M] of [section-total]: [status]
-CONTEXT: [✓ if CONTEXT.md exists | - if not]
-
-## Argument Strength
-- Core claim: [from STATE.md]
-- Evidence gaps: [list]
-
-## Key Decisions Made
-- [decision 1 from STATE.md]
-- [decision 2]
-
-## Open Issues
-- [any deferred issues or gaps]
+## 1. Validate Environment
 
-## What's Next
-[Next section/plan objective from ROADMAP]
+```bash
+ls .planning/STATE.md .planning/ROADMAP.md .planning/PROJECT.md 2>/dev/null
 ```
 
-</step>
-
-<step name="route">
-**Determine next action based on verified counts.**
+If missing `.planning/` → suggest `/wtfp:new-paper`.
+If missing STATE.md or ROADMAP.md → suggest what's needed.
 
-**Step 1: Count plans, summaries, and issues in current section**
+## 2. Load Project Context
 
-List files in the current section directory:
+Read: STATE.md, ROADMAP.md, PROJECT.md, config.json (for model_profile).
 
 ```bash
-ls -1 .planning/sections/[current-section-dir]/*-PLAN.md 2>/dev/null | wc -l
-ls -1 .planning/sections/[current-section-dir]/*-SUMMARY.md 2>/dev/null | wc -l
-ls -1 .planning/sections/[current-section-dir]/*-ISSUES.md 2>/dev/null | wc -l
-ls -1 .planning/sections/[current-section-dir]/*-REVISION.md 2>/dev/null | wc -l
+MODEL_PROFILE=$(cat .planning/config.json 2>/dev/null | grep -o '"model_profile"[[:space:]]*:[[:space:]]*"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' || echo "balanced")
 ```
 
-State: "This section has {X} plans, {Y} summaries, {Z} issues files."
-
-**Step 1.5: Check for unaddressed review issues**
-
-For each *-ISSUES.md file, check if matching *-REVISION.md exists.
-For each *-REVISION.md file, check if matching *-REVISION-SUMMARY.md exists.
+## 3. Gather Recent Work
 
-Track:
-- `issues_without_revision`: ISSUES.md files without REVISION.md
-- `revisions_without_summary`: REVISION.md files without REVISION-SUMMARY.md
+Find 2-3 most recent SUMMARY.md files. Extract: what was written, key decisions, issues.
 
-**Step 2: Route based on counts**
+## 4. Parse Current Position
 
-| Condition | Meaning | Action |
-|-----------|---------|--------|
-| revisions_without_summary > 0 | Unexecuted revision plans exist | Go to **Route A** (with REVISION.md) |
-| issues_without_revision > 0 | Review issues need revision plans | Go to **Route E** |
-| summaries < plans | Unexecuted plans exist | Go to **Route A** |
-| summaries = plans AND plans > 0 | Section complete | Go to Step 3 |
-| plans = 0 | Section not yet planned | Go to **Route B** |
+From STATE.md: current section, plan number, status, word count.
+Calculate: total plans, completed, remaining. Total words vs target.
+Check for CONTEXT.md in current section directory.
 
----
-
-**Route A: Unexecuted plan exists**
-
-Find the first PLAN.md without matching SUMMARY.md.
-Read its `<objective>` section.
+## 4b. Build Statusline
 
+Build the statusline string:
 ```
----
-
-## ▶ Next Up
-
-**{section}-{plan}: [Plan Name]** — [objective summary from PLAN.md]
-
-`/wtfp:write-section [full-path-to-PLAN.md]`
-
-<sub>`/clear` first → fresh context window</sub>
-
----
+WTF-P ► {section progress} ◆ {word count} ► {model_profile}
 ```
 
----
+**Section progress:** `{completed}/{total} sections` (count SUMMARY.md files vs total sections)
 
-**Route B: Section needs planning**
+**Word count:** `{current_words}/{target_words}w` (from STATE.md and PROJECT.md)
 
-Check if `{section}-CONTEXT.md` exists in section directory.
-
-**If CONTEXT.md exists:**
-
-```
----
-
-## ▶ Next Up
-
-**Section {N}: {Name}** — {Goal from ROADMAP.md}
-<sub>✓ Context gathered, ready to plan</sub>
-
-`/wtfp:plan-section {section-number}`
-
-<sub>`/clear` first → fresh context window</sub>
-
----
-```
-
-**If CONTEXT.md does NOT exist:**
+**Model profile:** from config.json (quality/balanced/budget)
 
+**Update indicator:** Check for newer version quietly:
+```bash
+CURRENT=$(node -e "console.log(require('$(npm root -g)/wtf-p/package.json').version)" 2>/dev/null || echo "")
+LATEST=$(npm view wtf-p version 2>/dev/null || echo "")
 ```
----
 
-## ▶ Next Up
+If LATEST exists and CURRENT < LATEST, append ` ▲ update` to statusline.
 
-**Section {N}: {Name}** — {Goal from ROADMAP.md}
-
-`/wtfp:plan-section {section}`
-
-<sub>`/clear` first → fresh context window</sub>
-
----
-
-**Also available:**
-- `/wtfp:discuss-section {section}` — gather context first
-- `/wtfp:research-gap {section}` — investigate literature
-- `/wtfp:list-assumptions {section}` — see Claude's assumptions
-
----
+**Pending todos:** Count files:
+```bash
+PENDING_TODOS=$(ls -1 .planning/todos/pending/*.md 2>/dev/null | wc -l | tr -d ' ')
 ```
 
----
-
-**Route E: Review issues need revision plans**
-
-ISSUES.md exists without matching REVISION.md. User needs to plan revisions.
+## 5. Present Status Report
 
+**Display statusline FIRST** (before all other output):
 ```
----
-
-## ⚠ Review Issues Found
-
-**{plan}-ISSUES.md** has {N} issues without a revision plan.
-
-`/wtfp:plan-revision {plan}`
-
-<sub>`/clear` first → fresh context window</sub>
-
----
-
-**Also available:**
-- `/wtfp:write-section [path]` — continue with other work first
-- `/wtfp:review-section {section}` — run more verification
-
----
+WTF-P ► 3/6 sections ◆ 4200/8000w ► balanced ▲ update
 ```
 
----
-
-**Step 3: Check paper status (only when section complete)**
-
-Read ROADMAP.md and identify:
-1. Current section number
-2. All section numbers in the document structure
-
-Count total sections and identify the highest section number.
-
-State: "Current section is {X}. Paper has {N} sections (highest: {Y})."
-
-**Route based on paper status:**
-
-| Condition | Meaning | Action |
-|-----------|---------|--------|
-| current section < highest section | More sections remain | Go to **Route C** |
-| current section = highest section | Paper draft complete | Go to **Route D** |
-
----
-
-**Route C: Section complete, more sections remain**
-
-Read ROADMAP.md to get the next section's name and goal.
+Then the detailed status report:
 
 ```
----
-
-## ✓ Section {Z} Complete
-
-## ▶ Next Up
+# [Paper Title]
 
-**Section {Z+1}: {Name}** — {Goal from ROADMAP.md}
+**Progress:** [████████░░] 8/10 sections drafted
+**Words:** [current] / [target] ([percentage]%)
+**Profile:** [quality/balanced/budget]
 
-`/wtfp:plan-section {Z+1}`
+## Recent Work
+- [Section X]: [1-line summary]
 
-<sub>`/clear` first → fresh context window</sub>
+## Current Position
+Section [N] of [total]: [name] — [status]
+CONTEXT: [✓ | -]
 
----
+## Key Decisions
+- [from STATE.md]
 
-**Also available:**
-- `/wtfp:review-section {Z}` — verify before continuing
-- `/wtfp:discuss-section {Z+1}` — gather context first
-- `/wtfp:research-gap {Z+1}` — investigate literature
+## Open Issues
+- [deferred items]
 
----
+## Pending Todos
+{PENDING_TODOS} pending (run /wtfp:check-todos to review)
 ```
 
----
-
-**Route D: Paper draft complete**
+## 6. Route to Next Action
 
+**Count files in current section:**
+```bash
+ls -1 .planning/sections/[dir]/*-PLAN.md 2>/dev/null | wc -l
+ls -1 .planning/sections/[dir]/*-SUMMARY.md 2>/dev/null | wc -l
+ls -1 .planning/sections/[dir]/*-ISSUES.md 2>/dev/null | wc -l
 ```
----
-
-## 🎉 Draft Complete
-
-All {N} sections written!
-
-**Words:** [current] / [target]
-
-## ▶ Next Up
 
-**Review & Polish** — verify and prepare for submission
+**Route A — Unexecuted plan exists** (summaries < plans):
+Find first PLAN.md without SUMMARY.md. Offer `/wtfp:write-section [path]`.
 
-`/wtfp:review-section` — run full verification
-`/wtfp:export-latex` — generate LaTeX output
+**Route B — Section needs planning** (plans = 0):
+If CONTEXT.md exists → offer `/wtfp:plan-section {N}`.
+If not → offer `/wtfp:discuss-section {N}` or `/wtfp:plan-section {N}`.
 
-<sub>`/clear` first → fresh context window</sub>
+**Route C — Section complete, more remain** (summaries = plans, not last section):
+Offer `/wtfp:plan-section {N+1}`. Also: `/wtfp:review-section {N}`.
 
----
-
-**Also available:**
-- `/wtfp:submit-milestone "draft-1"` — archive this draft
+**Route D — Draft complete** (all sections done):
+Offer `/wtfp:review-section` and `/wtfp:export-latex`.
 
----
-```
+**Route E — Review issues need revision** (ISSUES.md without REVISION.md):
+Offer `/wtfp:plan-revision {plan}`.
 
-</step>
+</process>
 
-<step name="edge_cases">
-**Handle edge cases:**
+<offer_next>
 
-- Section complete but next section not planned → offer `/wtfp:plan-section [next]`
-- All work complete → offer review and export
-- Gaps present → highlight before offering to continue
-- Handoff file exists → mention it, offer `/wtfp:resume-writing`
-</step>
+Show the routed next action with clear command and `/clear` suggestion.
 
-</process>
+</offer_next>
 
 <success_criteria>
-
+- [ ] Statusline displayed first: WTF-P ► {sections} ◆ {words} ► {profile} [▲ update]
+- [ ] Update indicator shown when newer version available
+- [ ] Pending todo count shown
 - [ ] Rich context provided (recent work, word counts, decisions)
 - [ ] Current position clear with visual progress
-- [ ] What's next clearly explained
-- [ ] Smart routing: /wtfp:write-section if plan exists, /wtfp:plan-section if not
+- [ ] Model profile shown
+- [ ] Smart routing to next command
 - [ ] User confirms before any action
-- [ ] Seamless handoff to appropriate wtfp command
 </success_criteria>