skills.rest
udecode
udecodeOFFICIAL
Explore all skills by this author →
WebsiteGitHub

gpt-pro

Generates self-contained external review prompts with source-backed repository context and pointed review questions.

16.5k|997|Updated Dec 2, 2019
Search Tags:#chatgpt#Prompt Engineering#architecture decision#external review#Code review#context packing#benchmark evidence
One-click install
npx skills add https://github.com/udecode/plate --skill gpt-pro
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: gpt-pro
Source: https://github.com/udecode/plate/tree/main/.agents/skills/gpt-pro
Command: npx skills add https://github.com/udecode/plate --skill gpt-pro

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

External reviewers like ChatGPT Pro have no access to your repository, terminal, or files, so naive review requests produce shallow or hallucinated feedback. This Skill builds a paste-ready prompt that embeds the full current state, evidence, and pointed questions so the external model can reason independently.

Core Features & Use Cases

  • Self-Contained Prompt Generation: Inlines source paths, API skeletons, behavior flows, benchmark results, and constraints so the reviewer needs zero repo access.
  • Evidence Labeling: Marks facts as confirmed, benchmarked, inferred, stale, or gap, and separates current truth from pasted prior answers.
  • Decision-Forcing Questions: Structures prompts to demand harsh verdicts, rejected alternatives, risk tables, red flags, and evidence that would overturn the recommendation.
  • Use Case: You are deciding on an API rewrite for the Plate editor and want a harsh external review. The Skill reads the current source, tests, and benchmarks, then produces a single paste-ready prompt containing the architecture skeleton, benchmark numbers, candidate directions, and the exact verdict you need.

Quick Start

Ask the assistant to create a gpt-pro review prompt for your current plan or architecture decision, naming the plan file or review target.

Frequently Asked Questions about gpt-pro

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get a useful code review from ChatGPT without repo access?▼

Include the full current state inline: source paths, API skeletons, behavior flow, benchmark results, and constraints. This Skill gathers that evidence from your local repo and packs it into one paste-ready prompt with pointed review questions.

What should an external review prompt include for architecture decisions?▼

Include the decision to make, current repo state, source-backed API skeleton, benchmark or test evidence, prior accepted decisions, constraints, known gaps, candidate directions, and the exact output requested. Ask for a harsh verdict, risk table, and red flags.

Can I reuse a previous ChatGPT answer in a new review prompt?▼

Yes, but treat it as context, not truth. The Skill includes prior answers as a labeled section and asks the reviewer to critique or refine them against freshly grounded repository state.

When should I not use an external review prompt workflow?▼

Skip it when you want local implementation, normal code review, or bug fixes, when a short chat answer suffices, or when the reviewer already has direct repository access.

How are stale benchmarks and missing evidence handled in the prompt?▼

Facts are labeled as confirmed, benchmarked, inferred, stale, or gap. If fresh proof is too expensive to gather, the prompt explicitly states which numbers are stale and asks the reviewer what evidence would change their verdict.

Related Skills

Explore More Software Engineering Skills→
mjamoremjamore

repo-docs-audit

Audit repository documentation against code and active plans to surface drift.

Community
Intermediate
FernanSuozaFernanSuoza

core-guided-user-workflow

Summarize repository state and recommend next steps across project phases.

Community
Advanced
secondskysecondsky
204

verification-before-completion

Require fresh verification evidence before claiming task completion.

Community
Basic
ohyesgocoolohyesgocool

mr-review

Orchestrate independent AI model code reviews on GitLab or GitHub merge requests.

Community
Advanced