- Home
- Skills
- Htooayelwinict
- Claude Config
- Project Planning
project-planning_skill
0
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill htooayelwinict/claude-config --skill project-planning- reference.md1.9 KB
- SKILL.md2.9 KB
Overview
This skill creates clear, step-by-step implementation plans with acceptance criteria, estimates, and a consolidated plan.md output. It is exclusive to the planner agent and uses MCP helpers for architecture analysis and documentation lookup to ground plans in codebase context. The output emphasizes independent, testable phases, risk-aware estimates, and concrete verification steps.
How this skill works
The skill inspects the codebase and documentation using MCP helpers (Gemini-Bridge/Open-Bridge for architecture analysis and Context7 for docs lookup) to find similar patterns, impacted files, and dependencies. It then produces a multi-phase plan with objectives, task lists, file actions, verification commands, estimates (T-shirt sizing and hours), and a summary plan.md ready for review. It annotates risks and applies risk multipliers for migrations, auth, and third-party integrations.
When to use it
- Planning a new feature end-to-end with deployable milestones
- Preparing a safe database migration or schema change
- Onboarding a cross-functional implementation with clear responsibilities
- Estimating effort and risks before sprint planning or stakeholder sign-off
- Creating a rollback and verification checklist for production releases
Best practices
- Start with one-sentence goal restatement to align scope
- Search the codebase for matching patterns and list affected files explicitly
- Break work into 3–5 independently testable phases with clear verification steps
- Use T-shirt sizing plus risk multipliers (DB/auth/third-party) for estimates
- Produce a single plan.md containing context, phases, file actions, commands, and summary
Example use cases
- Plan a new settings page with front-end and API changes, tests, and feature flag rollout
- Create a migration plan to add a column safely with migration scripts, backfill, and rollback
- Design a staged rollout for an auth change with feature flags and monitoring checkpoints
- Draft an integration plan for a third-party payment provider including configs and error handling
- Map work for converting a legacy endpoint to typed schemas with unit and integration tests
FAQ
You get a consolidated plan.md with context, code patterns, a phases table, detailed phase sections (tasks, files, verification), estimates, and a risk summary.
How are estimates calculated?
Estimates use T-shirt sizing mapped to hours and apply risk multipliers (1.5× for DB/auth, 2× for third-party) where applicable.
How many phases should a plan include?
Prefer 3–5 phases so each phase remains independently testable and verifiable; complex features may need more granular sub-tasks.