planflow-ai 1.1.4 → 1.1.5

package/.claude/commands/brain.md

@@ -48,12 +48,13 @@ MODES:
   Guided:     Asks structured questions about work, insights, meetings, decisions
 
 BRAIN STRUCTURE:
-  flow/brain/
-  ├── index.md        # Brain index (always loaded at session start)
-  ├── features/       # Feature history and context
-  ├── errors/         # Reusable error patterns
-  ├── decisions/      # Decision records
-  └── sessions/       # Daily activity logs
+  flow/brain/                          # Per-project brain
+  ├── index.md                         # Brain index (loaded at session start)
+  ├── features/                        # Feature history and context
+  └── errors/                          # Reusable error patterns
+
+  ~/plan-flow/brain/daily/             # Global daily logs (cross-project)
+  └── YYYY-MM-DD.md
 
 RELATED COMMANDS:
   /setup             Creates brain directory structure
@@ -69,7 +70,7 @@ RELATED COMMANDS:
 | ------------------------ | -------------------------------------------------------- |
 | **Brain Only**           | ONLY write to `flow/brain/` - no source code, no config  |
 | **Wiki-Links**           | All cross-references use `[[kebab-case-name]]` format    |
-| **Index Caps**           | Max 5 errors, 3 decisions, 3 cross-project in index      |
+| **Index Caps**           | Max 5 errors, 3 cross-project in index                    |
 | **No Code**              | Brain is for knowledge capture, never write source code   |
 | **Complete and Stop**    | After writing entry, STOP and wait for user               |
 
@@ -99,7 +100,7 @@ What would you like to capture? You can:
 Check if `flow/brain/` exists. If not, create:
 
 ```bash
-mkdir -p flow/brain/features flow/brain/errors flow/brain/decisions flow/brain/sessions
+mkdir -p flow/brain/features flow/brain/errors
 ```
 
 Check if `flow/brain/index.md` exists. If not, create initial index:
@@ -113,9 +114,6 @@ _No features tracked yet._
 ## Recent Errors
 _No errors captured yet._
 
-## Recent Decisions
-_No decisions recorded yet._
-
 ## Cross-Project Patterns
 _No cross-project patterns yet._
 ```

package/.claude/commands/heartbeat.md

@@ -0,0 +1,120 @@
+---
+description: Manage scheduled automated tasks via the heartbeat daemon
+---
+
+# Heartbeat: Scheduled Task Automation
+
+## Command Description
+
+Manage the heartbeat daemon — a cron-like automation system that runs scheduled tasks defined in `flow/heartbeat.md`. Tasks can include research, feature creation, execution, and pushing to repos.
+
+**Config**: `flow/heartbeat.md`
+**Daemon**: `npx plan-flow heartbeat start|stop|status`
+
+---
+
+## Help
+
+**If the user invokes this command with `-help`, display only this section and stop:**
+
+```
+/heartbeat - Scheduled Task Automation
+
+DESCRIPTION:
+  Manage the heartbeat daemon for automated scheduled tasks.
+  Tasks are defined in flow/heartbeat.md with human-readable schedules.
+
+USAGE:
+  /heartbeat -add         Add a new scheduled task interactively
+  /heartbeat -remove      Remove an existing scheduled task
+  /heartbeat -list        List all scheduled tasks with status
+  /heartbeat -help        Show this help
+
+CLI COMMANDS:
+  npx plan-flow heartbeat start    Start the daemon
+  npx plan-flow heartbeat stop     Stop the daemon
+  npx plan-flow heartbeat status   Show daemon status
+
+SCHEDULE SYNTAX:
+  daily at {HH:MM AM/PM}           Run once daily at the specified time
+  every {N} hours                  Run every N hours
+  every {N} minutes                Run every N minutes
+  weekly on {day} at {HH:MM}       Run weekly on a specific day
+
+EXAMPLE TASKS:
+  daily at 10:00 PM - research about topic X, create md files and add to brain
+  every 6 hours - check for new issues and update tasklist
+  weekly on Monday at 9:00 AM - generate weekly summary report
+
+RELATED COMMANDS:
+  /flow           Toggle autopilot mode
+  /brain          Manual brain entry
+```
+
+---
+
+## Critical Rules
+
+| Rule | Description |
+|------|-------------|
+| **File-Based Config** | All tasks are defined in `flow/heartbeat.md` |
+| **No Direct Execution** | This command manages tasks, not executes them — the daemon handles execution |
+| **Confirm Before Remove** | Always confirm before removing a scheduled task |
+| **Complete and Stop** | After adding/removing/listing, STOP and wait for user input |
+
+---
+
+## Instructions
+
+### /heartbeat -add
+
+1. Ask the user for:
+   - **Task name**: Short descriptive name (kebab-case)
+   - **Schedule**: Using human-readable syntax (see help)
+   - **Command**: What the daemon should execute
+   - **Description**: What this task does
+2. Add the task to `flow/heartbeat.md` using the task template
+3. Confirm the addition
+
+### /heartbeat -remove
+
+1. List all tasks in `flow/heartbeat.md`
+2. Ask the user which task to remove
+3. Confirm before removing
+4. Remove the task entry from the file
+
+### /heartbeat -list
+
+1. Read `flow/heartbeat.md`
+2. Display all tasks with their schedule, status, and description
+
+---
+
+## Context Optimization
+
+### Reference Codes for Heartbeat
+
+| Code | Description | When to Expand |
+|------|-------------|----------------|
+| COR-HB-1 | Heartbeat file format and schedule syntax | Creating or editing heartbeat tasks |
+| COR-HB-2 | Daemon architecture and CLI commands | Starting/stopping the daemon |
+
+---
+
+## Brain Capture
+
+<!-- brain-capture
+skill: heartbeat
+captures:
+  - task name and schedule added/removed
+  - daemon start/stop events
+-->
+
+---
+
+## Related Resources
+
+| Resource | Purpose |
+|----------|---------|
+| `resources/core/heartbeat.md` | File format rules and daemon architecture |
+| `resources/core/_index.md` | Index of core rules with reference codes |

package/.claude/commands/learn.md

@@ -1,34 +1,50 @@
 ---
-description: Extract reusable patterns from the current session and save them to the project
+description: Extract reusable patterns from the current session, or learn about a topic with step-by-step teaching
 ---
 
-# Learn: Session Pattern Extraction
+# Learn: Session Pattern Extraction & Teaching Mode
 
 ## Command Description
 
-This command analyzes the current session for reusable patterns — error resolutions, debugging techniques, workarounds, project-specific conventions — and saves them to `flow/resources/` as learned pattern files with `[[wiki-links]]` for Obsidian compatibility.
+This command has two modes:
 
-**Output**: Pattern files in `flow/resources/learned-{pattern-name}.md`.
+1. **Pattern Extraction** (`/learn`): Analyzes the current session for reusable patterns and saves them to `flow/resources/`.
+2. **Teaching Mode** (`/learn about <topic>`): Creates a structured step-by-step curriculum to teach you about a topic, contextualized to your project's tech stack.
+
+**Output**:
+- Pattern mode: `flow/resources/learned-{pattern-name}.md`
+- Teaching mode: `flow/brain/learning/{topic-kebab}.md`
 
 ---
 
+## Mode Detection
+
+Determine the mode based on user input:
+
+- `/learn` → **Pattern Extraction Mode** (existing behavior, see Steps 1-4 below)
+- `/learn about <topic>` → **Teaching Mode** (see Teaching Mode section below)
+- `/learn -help` → Show help and stop
+
+---
 
 ## Help
 
 **If the user invokes this command with `-help`, display only this section and stop:**
 
 ```
-/learn - Extract Reusable Patterns
+/learn - Extract Patterns & Teaching Mode
 
 DESCRIPTION:
-  Analyzes the current session for reusable patterns and saves them
-  to flow/resources/ as learned pattern files with [[wiki-links]].
+  Two modes: extract reusable patterns from the current session,
+  or learn about a topic with step-by-step teaching.
 
 USAGE:
   /learn                  Analyze session and suggest patterns
+  /learn about <topic>    Start teaching mode for a topic
   /learn -help            Show this help
 
-WHAT GETS EXTRACTED:
+PATTERN EXTRACTION MODE (/learn):
+  Extracts:
   - Error resolutions (how you fixed a tricky bug)
   - Debugging techniques (steps to diagnose an issue)
   - Workarounds (non-obvious solutions to platform/tool limitations)
@@ -36,16 +52,19 @@ WHAT GETS EXTRACTED:
   - Architecture insights (structural decisions made during work)
   - Integration patterns (how services/APIs were connected)
 
-WHAT DOES NOT GET EXTRACTED:
+  Does NOT extract:
   - Trivial fixes (typos, missing imports, obvious syntax errors)
   - One-time issues unlikely to recur
   - Generic knowledge available in documentation
-  - Patterns already captured in existing learned files
 
-OUTPUT:
-  - Saves to flow/resources/learned-{pattern-name}.md
-  - Creates [[wiki-links]] to related brain entries
-  - One pattern per file
+  Output: flow/resources/learned-{pattern-name}.md
+
+TEACHING MODE (/learn about <topic>):
+  Creates a structured curriculum (3-7 steps) contextualized
+  to your project's tech stack. Each step is confirmed before
+  proceeding. Completed curricula are saved to the brain.
+
+  Output: flow/brain/learning/{topic-kebab}.md
 
 RELATED COMMANDS:
   /brain           Capture meeting notes, ideas, brainstorms
@@ -160,6 +179,67 @@ If multiple patterns were found, repeat Steps 3-4 for each one.
 
 ---
 
+## Teaching Mode (`/learn about <topic>`)
+
+When the user invokes `/learn about <topic>`, switch to teaching mode:
+
+### Teaching Restrictions
+
+| Rule | Description |
+|------|-------------|
+| **Project Context** | Contextualize all examples to the project's actual tech stack |
+| **Step Confirmation** | Wait for user confirmation before proceeding to the next step |
+| **Brain Storage** | Save the curriculum to `flow/brain/learning/{topic-kebab}.md` |
+| **No Code Changes** | Teaching mode does NOT write source code or modify configs |
+| **Complete and Stop** | After all steps are confirmed, STOP and wait for user input |
+
+### Teaching Workflow
+
+1. **Analyze Context**: Read `flow/brain/index.md` and `flow/references/tech-foundation.md` to understand the project's stack
+2. **Generate Curriculum**: Create a 3-7 step curriculum for the topic, contextualized to the project
+3. **Present Overview**: Show the full curriculum outline to the user for approval
+4. **Step-by-Step Teaching**: For each step:
+   - Present the step content with explanations and examples
+   - Wait for user confirmation (`next`, `done`, or questions)
+   - Mark the step as completed in the curriculum file
+   - If the user asks questions, answer them before proceeding
+5. **Save Curriculum**: Write the completed curriculum to `flow/brain/learning/{topic-kebab}.md`
+
+### Curriculum Template
+
+```markdown
+# Learning: {Topic}
+
+**Project**: [[{project-name}]]
+**Started**: {YYYY-MM-DD}
+**Status**: {in-progress|completed}
+
+## Curriculum
+
+### Step 1: {Title}
+- **Status**: {pending|completed}
+- **Content**: {explanation with project-contextualized examples}
+
+### Step 2: {Title}
+- **Status**: {pending|completed}
+- **Content**: {explanation}
+
+...
+
+## Notes
+
+{Any additional notes, questions asked, or insights gained during learning}
+```
+
+### Reference Codes for Teaching
+
+| Code | Description | When to Expand |
+|------|-------------|----------------|
+| SKL-LRN-3 | Teaching mode restrictions and workflow | Starting teaching mode |
+| SKL-LRN-4 | Curriculum template and brain storage | Generating curriculum |
+
+---
+
 ## Flow Diagram
 
 ```

package/.claude/commands/setup.md

@@ -399,7 +399,7 @@ See: `.claude/resources/skills/setup-skill.md`
 If the `flow/` folder doesn't exist, create it:
 
 ```bash
-mkdir -p flow/archive flow/brain/features flow/brain/errors flow/brain/decisions flow/brain/sessions flow/contracts flow/discovery flow/plans flow/references flow/reviewed-code flow/reviewed-pr
+mkdir -p flow/archive flow/brain/features flow/brain/errors flow/contracts flow/discovery flow/plans flow/references flow/reviewed-code flow/reviewed-pr
 ```
 
 **Add `.gitkeep` files**:
@@ -416,7 +416,6 @@ flow/
 ├── brain/             # Automatic knowledge capture (Obsidian-compatible)
 │   ├── features/      # Feature history and context
 │   ├── errors/        # Reusable error patterns
-│   ├── decisions/     # Decision records
 │   └── sessions/      # Daily activity logs
 ├── contracts/         # Integration contracts
 ├── discovery/         # Discovery documents
@@ -454,7 +453,6 @@ flow/
 ├── brain/             # Automatic knowledge capture (Obsidian-compatible)
 │   ├── features/      # Feature history and context
 │   ├── errors/        # Reusable error patterns
-│   ├── decisions/     # Decision records
 │   └── sessions/      # Daily activity logs
 ├── contracts/         # Integration contracts
 ├── discovery/         # Discovery documents
@@ -711,6 +709,39 @@ When executing this command:
 
 ---
 
+## Global Pattern Sync
+
+After generating pattern files in Step 8, sync **generic** patterns to the global brain at `~/plan-flow/brain/patterns/`.
+
+### What to Sync
+
+| Local Pattern File | Global Target | Sync? |
+|---|---|---|
+| `.claude/resources/frameworks/<fw>-patterns.md` | `~/plan-flow/brain/patterns/<fw>.md` | Yes — framework best practices are generic |
+| `.claude/resources/libraries/<lib>-patterns.md` | `~/plan-flow/brain/patterns/<lib>.md` | Yes — library patterns are generic |
+| `.claude/resources/project/project-patterns.md` | — | No — project-specific paths and structure |
+
+### Sync Process
+
+For each framework and library pattern file generated:
+
+1. Check if `~/plan-flow/brain/patterns/<name>.md` already exists
+2. **If it does NOT exist**:
+   - Copy the pattern file, but strip project-specific file paths and code examples that reference specific project directories
+   - Keep generic best practices, conventions, allowed/forbidden patterns
+   - Add a `## Projects Using This` section at the bottom: `- [[project-name]]`
+3. **If it ALREADY exists**:
+   - Read the existing global file
+   - Append any NEW sections or patterns not already present (do NOT overwrite existing content — it may contain knowledge from other projects)
+   - Add `[[project-name]]` to the `## Projects Using This` section if not already listed
+4. Use kebab-case for file names: `nextjs.md`, `typescript.md`, `zod.md`, `react-query.md`
+
+### Example
+
+If setup generates `.claude/resources/frameworks/nextjs-patterns.md`, create or update `~/plan-flow/brain/patterns/nextjs.md` with the generic patterns, linked to the project via `[[project-name]]`.
+
+---
+
 ## Brain Capture
 
 After setup completes successfully, append a brain-capture block to the session file. See `.claude/resources/core/brain-capture.md` for processing rules.
@@ -726,11 +757,12 @@ data:
   project_name: [detected project name]
   stack: [language + framework]
   patterns_generated: [count of pattern files created]
+  patterns_synced_to_global: [list of global pattern files created/updated]
   files_created: [list of generated files]
 -->
 ```
 
-Write to `flow/brain/sessions/YYYY-MM-DD.md` and update `flow/brain/index.md` if this is a new project setup.
+Write to `~/plan-flow/brain/daily/YYYY-MM-DD.md` and update `flow/brain/index.md` if this is a new project setup.
 
 ---
 

package/.claude/resources/core/_index.md

@@ -5,8 +5,8 @@
 
 Core rules define the foundational coding standards that apply across the entire project. These include best practices to follow (allowed patterns), anti-patterns to avoid (forbidden patterns), and complexity scoring for implementation planning.
 
-**Total Files**: 7 files, ~1300 lines
-**Reference Codes**: COR-AP-1 through COR-RC-2
+**Total Files**: 10 files, ~1560 lines
+**Reference Codes**: COR-AP-1 through COR-HB-2
 
 ---
 
@@ -59,6 +59,27 @@ Core rules define the foundational coding standards that apply across the entire
 | COR-BR-2 | File templates (feature, error, decision, session) and naming conventions | brain-capture.md | 52-170 |
 | COR-BR-3 | Index management (caps, rotation) and global brain sync | brain-capture.md | 172-240 |
 
+### Project Memory (`project-memory.md`)
+
+| Code | Description | Source | Lines |
+|------|-------------|--------|-------|
+| COR-MEM-1 | Memory session start behavior and entry format | project-memory.md | 10-50 |
+| COR-MEM-2 | Field descriptions and maintenance rules | project-memory.md | 52-75 |
+
+### Project Tasklist (`project-tasklist.md`)
+
+| Code | Description | Source | Lines |
+|------|-------------|--------|-------|
+| COR-TL-1 | Tasklist session start behavior and task format | project-tasklist.md | 10-45 |
+| COR-TL-2 | Brain integration and maintenance rules | project-tasklist.md | 47-65 |
+
+### Heartbeat (`heartbeat.md`)
+
+| Code | Description | Source | Lines |
+|------|-------------|--------|-------|
+| COR-HB-1 | Heartbeat file format and schedule syntax | heartbeat.md | 10-55 |
+| COR-HB-2 | Daemon architecture, CLI commands, and rules | heartbeat.md | 57-110 |
+
 ### Resource Capture (`resource-capture.md`)
 
 | Code | Description | Source | Lines |
@@ -89,6 +110,12 @@ Core rules define the foundational coding standards that apply across the entire
 | COR-BR-1 | Need brain session start behavior and processing rules |
 | COR-BR-2 | Need brain file templates and naming conventions |
 | COR-BR-3 | Need index management caps and global sync rules |
+| COR-MEM-1 | Need memory session start behavior and entry format |
+| COR-MEM-2 | Need memory field descriptions and maintenance rules |
+| COR-TL-1 | Need tasklist session start behavior and task format |
+| COR-TL-2 | Need tasklist brain integration and maintenance rules |
+| COR-HB-1 | Need heartbeat file format and schedule syntax |
+| COR-HB-2 | Need daemon architecture and CLI commands |
 | COR-RC-1 | Need resource capture behavior and criteria |
 | COR-RC-2 | Need resource file format and naming conventions |
 
@@ -127,3 +154,6 @@ Core rules define the foundational coding standards that apply across the entire
 - `autopilot-mode.md` has `alwaysApply: true` - detects and orchestrates autopilot flow mode
 - `brain-capture.md` is loaded on-demand - processes brain-capture blocks and manages brain index
 - `resource-capture.md` is loaded on-demand - watches for valuable reference materials during skill execution
+- `project-tasklist.md` is loaded on-demand - session start tasklist behavior
+- `project-memory.md` is loaded on-demand - artifact tracking and 7-day session loading
+- `heartbeat.md` is loaded on-demand - scheduled task daemon configuration

package/.claude/resources/core/autopilot-mode.md

@@ -11,6 +11,8 @@ When `flow/.autopilot` exists, this rule is active. It automatically orchestrate
 1. Check if `flow/.autopilot` exists
 2. If it exists: autopilot mode is **ON** - follow the rules below for every user input
 3. If it does not exist: autopilot mode is **OFF** - ignore this rule entirely, normal behavior applies
+4. Load `flow/tasklist.md` if it exists — summarize active tasks and ask if the user wants to pick one
+5. Load `flow/memory.md` if it exists — internalize last 7 days of completed work silently
 
 ---
 
@@ -105,8 +107,10 @@ Check `flow/contracts/` for any relevant integration contracts. If found, note t
 
 1. Move the discovery document to `flow/archive/`
 2. Move the plan document to `flow/archive/`
-3. Present completion summary
-4. **Prompt for context cleanup** (see below)
+3. Update `flow/tasklist.md` — mark the completed task as **Done** with today's date
+4. Append a memory entry to `flow/memory.md` with the completed feature summary
+5. Present completion summary
+5. **Prompt for context cleanup** (see below)
 
 ---