ReddRadar
Agent skill
planning
Comprehensive feature planning and execution system. Provides workflows for creating plans, decomposing into features, and pull-based execution. Uses Beads as source of truth for state management.
Install this agent skill to your Project
npx add-skill https://github.com/NotMyself/planning-system/tree/main/skills/planning
SKILL.md
Planning System Skill
Overview
This skill provides a reliable system for planning and executing features. It uses Beads as the source of truth for plans and task state, with mechanical enforcement of quality gates.
Architecture
Standard Planning (single feature)
plan:new ──► Beads Epic (plan in description)
│
▼
plan:optimize <epic-id> ──► Beads Tasks (prompts in descriptions)
│ + dev/plans/<name>/ (supporting files)
▼
bd ready ──► Pull-based execution (one task at a time)
Master Planning (large initiative)
plan:new --master ──► Beads Epic + Features (high-level decomposition)
│
▼
plan:new <feature-id> ──► Detailed plan in feature description
│ (optional: research tasks first)
▼
plan:optimize <feature-id> ──► Beads Tasks under feature
│ + dev/plans/<name>/ (supporting files)
▼
bd ready ──► Pull-based execution (one task at a time)
Key Components
| Component | Purpose |
|---|---|
| Commands | Lean workflow coordinators |
| Workflows | Detailed phase instructions (loaded on-demand) |
| Templates | Standardized file formats |
| Verification | TypeScript scripts for mechanical enforcement |
| Hooks | Block shortcuts at Stop events |
Beads Integration (Required)
All plans are stored in Beads epics. Each feature becomes a task under that epic with the full prompt in the description.
# Epic created by plan:new (plan content in description)
bd create --type=epic --title="Feature Name" --description="<full plan>" --silent
# Tasks created by plan:optimize (prompts in description)
bd create --type=task --parent=<epic-id> --title="F001: Setup types" --description="<full prompt>" --silent
# Dependencies between features
bd dep add <task-F002> <task-F001>
# Pull-based execution
bd ready # Find available work
bd update <task-id> --status=in_progress # Claim work
bd close <task-id> # Complete work
DevOps Integration (Optional)
Azure DevOps Story status can sync with Beads epic via sync-devops.ts.
Verification System
Hooks in hooks/hooks.json run TypeScript verification scripts:
- Stop hook: Blocks completion if quality gates fail
Quality gates (all configurable via .planconfig):
- No uncommitted changes
- Build command succeeds
- Test command succeeds
- Lint command succeeds
- Format command succeeds
- Static analysis succeeds
- Verification agent approves test quality and requirements
Quick Reference
| Task | Command |
|---|---|
| Start new plan (standard) | plan:new |
| Start master plan | plan:new --master |
| Plan a feature | plan:new <feature-id> |
| Optimize epic into tasks | plan:optimize <epic-id> |
| Optimize feature into tasks | plan:optimize <feature-id> |
| Find work | bd ready |
| View task | bd show <task-id> |
| Claim work | bd update <task-id> --status=in_progress |
| Complete work | bd close <task-id> |
| Reflect on conversation | /reflect |
| Toggle auto-reflect | /reflect on or /reflect off |
| Reflect status | /reflect status |
Workflow Files
Load these on-demand during command execution:
workflows/1-planning.md- Planning phase detailsworkflows/2-optimization.md- Optimization phase details
Templates
templates/plan.md- Plan document structuretemplates/prompt.md- Feature prompt structuretemplates/context.md- Context file templatetemplates/constraints.md- Constraints file template
Supporting Files
Created by plan:optimize in dev/plans/<name>/:
context.md- Project background and architectureconstraints.md- Global rules for all featuresdecisions.md- Architectural decisions logedge-cases.md- Known edge cases and handlingtesting-strategy.md- Testing approach
Project Configuration
.planconfig (YAML, project root):
build_command: "npm run build"
test_command: "npm test"
lint_command: "eslint ."
format_command: "prettier --check ."
static_analysis_command: "sonar-scanner"
All commands optional. Missing commands skip that verification step.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
davila7/claude-code-templates
verl-rl-training
Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.
davila7/claude-code-templates
openrlhf-training
High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.
davila7/claude-code-templates
gguf-quantization
GGUF format and llama.cpp quantization for efficient CPU/GPU inference. Use when deploying models on consumer hardware, Apple Silicon, or when needing flexible quantization from 2-8 bit without GPU requirements.
davila7/claude-code-templates
Claude Code Guide
Master guide for using Claude Code effectively. Includes configuration templates, prompting strategies "Thinking" keywords, debugging techniques, and best practices for interacting with the agent.
davila7/claude-code-templates
qdrant-vector-search
High-performance vector similarity search engine for RAG and semantic search. Use when building production RAG systems requiring fast nearest neighbor search, hybrid search with filtering, or scalable vector storage with Rust-powered performance.
davila7/claude-code-templates
behavioral-modes
AI operational modes (brainstorm, implement, debug, review, teach, ship, orchestrate). Use to adapt behavior based on task type.
Didn't find tool you were looking for?