implementation-process
Scannednpx machina-cli add skill serpro69/claude-starter-kit/implementation-process --openclawExecuting Plans
Overview
Load plan, review critically, execute tasks in batches, report for review between batches.
Core principle: Batch execution with checkpoints for architect review.
The Process
Step 1: Load and Review Plan
- Read plan file
- Review critically - identify any questions or concerns about the plan
- If concerns: Raise them with your human partner before starting
- If no concerns: Get the task from taskmaster
Step 2: Execute Sub-Task
- Update status to in-progress
- Follow the plan exactly
- Run verifications as specified; use
testing-processskill
Step 3: Report
- Show what was implemented
- Show verification output
- Prompt user for code-review; if user responds 'yes' - run
zenmcp code-review with gemini-3-pro - Based on user and code-review feedback: apply changes if needed and finalize the sub-task
- When completed mark as done
Step 4: Continue
- Execute next sub-task
- Repeat until all sub-tasks are completed
Step 5: Complete Development
After all sub-tasks complete and verified:
- Use
testing-processskill to verify and validate functionality - Use
documentation-processskill to create or update any relevant docs
When to Stop and Ask for Help
STOP executing immediately when:
- Hit a blocker mid-batch (missing dependency, test fails, instruction unclear)
- Plan has critical gaps preventing starting
- You don't understand an instruction
- Verification fails repeatedly
IMPORTANT! Always ask for clarification rather than guessing.
When to Revisit Earlier Steps
Return to Review (Step 1) when:
- Partner updates the plan based on your feedback
- Fundamental approach needs rethinking
IMPORTANT! Don't force through blockers - stop and ask.
Remember
- Review plan critically first
- Follow plan steps exactly
- Don't skip verifications
- Use skills when the plan says to do so
- Between batches: just report and wait
- Stop when blocked, don't guess
Source
git clone https://github.com/serpro69/claude-starter-kit/blob/master/.github/templates/claude/skills/implementation-process/SKILL.mdView on GitHub Overview
This skill guides implementing a feature strictly from a written plan. It emphasizes loading the plan, critically reviewing it, executing work in batches, and reporting between batches for architect review.
How This Skill Works
You load the plan and review it for questions. Then you execute sub-tasks in batches, updating status to in-progress and performing the verifications specified (via the testing-process skill). After each batch, you report results, prompt for code-review, apply feedback, and proceed to the next sub-task, finishing with validation and docs.
When to Use It
- You have a fully-formed written implementation plan to execute in a separate session with review checkpoints
- You require batch execution with architect review between batches
- You need to perform verifications after each sub-task using the testing-process skill
- You must stop and ask for clarification when blockers or unclear instructions occur
- You will finalize development by validating functionality and updating docs with documentation-process
Quick Start
- Step 1: Load and critically review the plan; raise concerns if any
- Step 2: Execute the first sub-task, update status to in-progress, and run verifications with testing-process
- Step 3: Report results, solicit code-review if requested, apply feedback, and proceed to the next sub-task
Best Practices
- Review the plan critically before starting
- Follow plan steps exactly as written
- Run the specified verifications after each sub-task
- Use testing-process for verifications and code-review prompts when applicable
- Between batches, report progress and wait for feedback
Example Use Cases
- Implementing a new feature from a complete plan with architect checkpoints between batches
- API surface changes executed in batch steps with formal reviews
- Running unit/integration verifications via testing-process after each sub-task
- Receiving code-review prompts and applying changes before finalizing a sub-task
- Updating or creating documentation after completing all sub-tasks using documentation-process
