plan-implementation
Plan Implementation
Methodology for decomposing work into implementation units and planning each one. Use this skill standalone or as a reference from other planning skills.
Decomposition
Sizing
- One unit = one logical unit of work (a feature, a subsystem, a layer)
- Never split tightly-coupled pieces across units (if UI + API + tests are inseparable, keep them together)
- Split independent subsystems into separate units
- If a unit would touch more than ~15-20 files or span 3+ unrelated subsystems, split further
- If the entire scope fits one session, produce a single unit — do not over-decompose
- Each unit must leave the codebase fully integrated — never leave components unreachable from the project's entry points
Ordering
Order by dependency — foundational work before dependent work:
- Setup and scaffolding (project init, config, CI)
- Data and domain layer (models, schemas, types)
- Core business logic
- API and service layer
- UI and frontend
- Integration and end-to-end concerns
- Finalization (QA, commit, PR — always last)
Adaptation
When a plan meets reality, adjust for divergence before executing:
- File paths changed — update references to match actual structure
- Architecture evolved — adjust the plan to work with what was actually built
- Scope shifted — add or remove items based on what prior work produced
- New information — incorporate discoveries from previous sessions
Keep adaptations minimal — adjust only what diverged, do not rewrite plans unnecessarily.
Planning Each Unit
For each unit of work, plan the implementation thoroughly:
- Explore the codebase — read files relevant to the unit's scope, understand existing patterns, find code to build on
- Design the implementation — determine what files to create or modify, what approach to take, what tests to write
- Write a detailed plan — specific enough to execute: exact file paths, function signatures, data flow, test cases
- Include a task list — one task per step in the plan
- Include relevant skills — identify which skills apply and instruct: "After plan approval and before making edits, run
/skill-a,/skill-b."
Rules
- Split conservatively when scope is ambiguous (smaller units are safer than oversized ones)
- Each unit must be self-contained with enough context to understand the work without reading the full plan
- Never merge setup and finalization into the same unit
More from tobihagemann/turbo
find-dead-code
Find dead code using parallel subagent analysis and optional CLI tools, treating code only referenced from tests as dead. Use when the user asks to \"find dead code\", \"find unused code\", \"find unused exports\", \"find unreferenced functions\", \"clean up dead code\", or \"what code is unused\". Analysis-only — does not modify or delete code.
30simplify-code
Run a multi-agent review of changed files for reuse, quality, efficiency, and clarity issues followed by automated fixes. Use when the user asks to \"simplify code\", \"review changed code\", \"check for code reuse\", \"review code quality\", \"review efficiency\", \"simplify changes\", \"clean up code\", \"refactor changes\", or \"run simplify\".
23smoke-test
Launch the app and hands-on verify that it works by interacting with it. Use when the user asks to \"smoke test\", \"test it manually\", \"verify it works\", \"try it out\", \"run a smoke test\", \"check it in the browser\", or \"does it actually work\". Not for unit/integration tests.
22finalize
Run the post-implementation quality assurance workflow including tests, code polishing, review, and commit. Use when the user asks to \"finalize implementation\", \"finalize changes\", \"wrap up implementation\", \"finish up\", \"ready to commit\", or \"run QA workflow\".
22self-improve
Extract lessons from the current session and route them to the appropriate knowledge layer (project AGENTS.md, auto memory, existing skills, or new skills). Use when the user asks to \"self-improve\", \"distill this session\", \"save learnings\", \"update CLAUDE.md with what we learned\", \"capture session insights\", \"remember this for next time\", \"extract lessons\", \"update skills from session\", or \"what did we learn\".
22evaluate-findings
Critically assess external feedback (code reviews, AI reviewers, PR comments) and decide which suggestions to apply using adversarial verification. Use when the user asks to \"evaluate findings\", \"assess review comments\", \"triage review feedback\", \"evaluate review output\", or \"filter false positives\".
22