Agent skill
plan-readiness-review
Use when you have a plan and want to verify it's ready for implementation - catches structural gaps, feasibility issues, and scope problems
Stars
163
Forks
31
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/plan-readiness-review
SKILL.md
Plan Readiness Review
Overview
Pressure test whether a plan is ready for implementation. Catches structural gaps, feasibility issues, and scope problems before committing to build.
Core principle: Find problems before implementation, not during.
Announce at start: "I'm using the plan-readiness-review skill to evaluate this plan."
Plan location: plans/active/{plan-name}/
The Process
Step 1: Load Plan
- Read all files from
plans/active/{plan-name}/design/ - Read all files from
plans/active/{plan-name}/implementation/(if exists) - Note: Plan must be committed to repo before worktree creation
Step 2: Evaluate Against Criteria
Structure:
- All required sections present (problem, solution, tasks)?
- Steps specific enough to execute without interpretation?
- Success criteria defined?
- Dependencies between tasks identified?
Feasibility:
- Technical approach sound?
- Required APIs/libraries exist?
- No impossible constraints?
Scope:
- Right-sized for one branch?
- Should it be split into multiple plans?
- Over-engineered for the problem?
- Missing obvious requirements?
Clarity:
- Would a fresh agent understand every instruction?
- Any ambiguous instructions?
- Missing context that implementer would need?
Step 3: Cross-Reference Codebase
Check if plan accounts for:
- Existing patterns in the codebase
- Files that will be affected
- Test patterns to follow
- Related code that might need updates
Step 4: Produce Verdict
Output this format:
## Plan Readiness Review: {plan-name}
### Verdict: READY | NEEDS WORK | SPLIT RECOMMENDED
### Strengths
- [What's solid about this plan]
### Issues Found
- **[Category]**: [Issue description]
- Suggested fix: [How to address]
### Questions to Resolve
- [Anything ambiguous that needs human input]
### Recommendation
[1-2 sentence summary of what to do next]
Verdict Meanings
| Verdict | Meaning | Next Step |
|---|---|---|
| READY | Plan is executable as-is | Proceed to implementation |
| NEEDS WORK | Fixable issues found | Address issues, re-review |
| SPLIT RECOMMENDED | Too large for one branch | Create multiple plans |
Red Flags
Never:
- Rubber-stamp a plan without thorough review
- Skip codebase cross-reference
- Approve plans with ambiguous instructions
Always:
- Check every evaluation criterion
- Verify file paths exist or make sense
- Flag anything a fresh agent might misinterpret
Integration
Invoked by:
- gremlins:worktree-workflow (start phase) - First gate in workflow
- Standalone when reviewing any plan
Pairs with:
- gremlins:writing-plans - Reviews output of that skill
- gremlins:executing-plans - Feeds into that skill when READY
Didn't find tool you were looking for?