review-plan
Validates project plans against codebases and standards to catch architectural mistakes early.
Install
mkdir -p .claude/skills/review-plan-hawkyre && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/19446" && unzip -o skill.zip -d .claude/skills/review-plan-hawkyre && rm skill.zipInstalls to .claude/skills/review-plan-hawkyre
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Stress-test a plan file before committing to implementation. Catches architectural mistakes, convention violations, and missing dependencies while changes are still cheap. Invoked directly, not routed from coding-process.Key capabilities
- →Analyze plan files for technical soundness
- →Verify dependency ordering between increments
- →Check plan compliance with project standards
- →Validate plan fidelity against conversation history
- →Assess increment implementability and done criteria
How it works
The skill reads plan files and associated notes to evaluate architectural decisions, dependency sequences, and convention adherence. It uses big-output discipline to process large logs and file contents efficiently.
Inputs & outputs
When to use review-plan
- →Validate multi-PR architectural plan files
- →Check dependency ordering for code increments
- →Ensure planned changes comply with project standards
- →Verify plan fidelity against recent chat history
About this skill
Review a plan
Read the requested plan. Otherwise find .plans/<slug>/plan.md; ask if the task does not identify a unique plan.
Read linked use cases and technical notes, plus relevant standards and common mistakes.
For an older plan, read its existing format. Do not review a published copy when the source document exists.
Check:
- The design solves the class of problems at its source, not only this instance. Do not demand generality for a class with no second member.
- The design fits the code and the stated constraints.
- Each use case works through the UI, server, and data where relevant.
- Important failure, permission, and recovery paths have defined results.
- The work covers the use cases. PR dependencies exist, contain no cycles, and permit safe merges.
- Each task has an observable check. Plain language is enough.
- Data changes address constraints, compatibility, backfill, and rollback where needed.
- The plan matches the conversation's scope and decisions, if available.
- Each fact appears once. Linked detail adds information instead of repeating the main plan.
A small plan has one work list. A large plan has PR blocks. Execution records live outside the plan.
Short use cases belong in the main plan. Detailed flows can live in a linked use-cases.md; that file is optional.
Do not demand empty sections, fixed word counts, source tags on every claim, or a quota of assumptions.
Report only supported findings. Name the location, consequence, and correction. Use MUST-FIX for correctness or implementation blockers, SHOULD-FIX for concrete friction or a better approach with a named benefit, and CONSIDER for real tradeoffs where the user must choose. A plan that works is not exempt. If a better approach exists, report it. Say when a category has no findings. Do not invent requirements to fill it. Do not edit the plan unless the user asks for fixes.
When not to use it
- →Routing to coding-process
- →Reviewing plans without conversation context when fidelity is required
Limitations
- →Requires specific file paths like .plans/<slug>/plan.html
- →Heavy command output must be redirected to /tmp/hawk-<skill>-<step>.log
How it compares
Unlike manual review, this skill performs an adversarial check against project standards and dependency logic before any code is written.
Compared to similar skills
review-plan side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| review-plan (this skill) | 0 | 4mo | No flags | Intermediate |
| confidence-check | 13 | 7mo | Review | Beginner |
| specification-architect | 13 | 11mo | Review | Advanced |
| drift-analysis | 2 | 7mo | No flags | Intermediate |
Try saying
Example prompts that trigger this skill in your AI assistant.
You might also like
confidence-check
SuperClaude-Org
Pre-implementation confidence assessment (≥90% required). Use before starting any implementation to verify readiness with duplicate check, architecture compliance, official docs verification, OSS references, and root cause identification.
specification-architect
adrianpuiu
A rigorous, traceability-first system that generates five interconnected architectural documents (blueprint.md, requirements.md, design.md, tasks.md, and validation.md) with complete requirements-to-implementation traceability. Use this skill when users need to architect systems, create technical specifications, or develop structured project documentation with guaranteed traceability.
drift-analysis
avifenesh
This skill should be used when the user asks about "plan drift", "reality check", "comparing docs to code", "project state analysis", "roadmap alignment", "implementation gaps", or needs guidance on identifying discrepancies between documented plans and actual implementation state.
flow-next-plan-review
gmickel
Carmack-level plan review via RepoPrompt or Codex. Use when reviewing Flow epic specs or design docs. Triggers on /flow-next:plan-review.
rfc-check
meta-pytorch
Determine if proposed changes require an RFC. Use when planning significant changes, before starting major work, or when asked whether an RFC is needed.
feature-planning
mhattingpete
Break down feature requests into detailed, implementable plans with clear tasks. Use when user requests a new feature, enhancement, or complex change.