platonic-code-review_skill

This skill validates code against RFCs and design specs, generating a comprehensive report of gaps and inconsistencies without changing the code.

3

GitHub Stars

1

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 caesar0301/platonic-coding-skills --skill platonic-code-review

SKILL.md7.6 KB

Overview

This skill reviews code implementations against specifications to verify consistency, completeness, and correctness. It generates structured, evidence-backed reports highlighting gaps, discrepancies, and prioritized recommendations. By default it reports findings without modifying code; code changes only occur with explicit user consent.

How this skill works

The skill reads provided specification documents (RFCs, requirements, design docs) and extracts a testable checklist of features, behaviors, APIs, data structures, and constraints. It maps each checklist item to code locations, evaluates status (implemented/partial/missing/unclear), documents evidence with file references, and produces a prioritized report with suggested actions. Advanced modes enable deeper verification, cross-reference checks, and test-coverage analysis.

When to use it

Verify that an implementation matches an RFC or formal spec
Validate feature completeness before release or QA
Audit code for discrepancies against documented design decisions
Perform gap analysis between specs and source tree
Confirm test coverage aligns with specification requirements

Best practices

Start by reading and extracting requirements from specs before inspecting code
Create a clear, testable functionality checklist from the spec
Map checklist items to exact code files and line references
Prioritize critical inconsistencies and provide actionable, evidence-based recommendations
Do not change code without explicit user instruction; ask before implementing fixes

Example use cases

Review rfc-001-user-authentication.md against src/auth/ and list missing behaviors and edge cases
Conduct a comprehensive API audit comparing all RFCs in ./specs/ to src/ endpoints and data models
Perform a gap analysis to find specs with no implementation and code features not documented in specs
Validate that payment-processing requirements in rfc-005-payment-processing.md are fully implemented in src/payments/ and identify test gaps

FAQ

No. By default it produces a report only. Code modifications happen only if you explicitly request them.

What evidence does the report include?

Reports include an executive summary, a functionality checklist with status, detailed findings with specific file and line references, and prioritized recommendations.