full

Reviews current branch changes with parallel multi-agent analysis and validated findings.

15|1|Updated Jan 3, 2026
One-click install
npx skills add https://github.com/fprochazka/claude-code-plugins --skill full-fprochazka
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: full
Source: https://github.com/fprochazka/claude-code-plugins/tree/main/plugins/code-review/skills/full
Command: npx skills add https://github.com/fprochazka/claude-code-plugins --skill full-fprochazka

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Reviewing a branch by hand means juggling the diff, the MR/PR discussion, the ticket, project conventions, and many quality dimensions at once, which leads to shallow or inconsistent reviews. This Skill orchestrates a structured, context-aware review of every change on the current branch against its base branch. ## Core Features & Use Cases - Context-first review: Loads MR/PR description, target branch, comment threads, pipeline status, and ticket context before forming any opinion, so findings are grounded in intent. - Parallel multi-agent analysis: Launches exploration and conventions-mapping subagents, then up to nine specialized review agents (conventions, architecture, code design, bugs, performance, security, release, git history, docs) proportionally to what the diff actually touches. - Validated, deduplicated findings: Every agent finding is cross-checked against the actual code through six disproof grounds, then reported with a four-level severity vocabulary and a coverage section, including migration safety verdicts. - Use Case: Before merging a feature branch with a database migration, run the review to get a report with blocking issues, migration safety assessment, unresolved MR discussion status, and mermaid diagrams for the hardest-to-explain findings. ## Quick Start Ask the assistant to run a full code review of the current branch, optionally naming a focus area such as security or performance.

Frequently Asked Questions about full

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

FAQPage Schema
How do I review all changes on a branch before merging?▼

Run the full branch review, which loads the MR/PR context and ticket, diffs the branch against its target branch, and launches parallel review agents covering bugs, security, performance, and more. Findings are validated against the actual code and written to a report file.

What does a multi-agent code review check?▼

It checks nine dimensions: conventions, architecture, code design, bugs, performance, security, release readiness, git history, and documentation. Agents are selected proportionally to the diff, so a docs-only change skips irrelevant dimensions like performance.

Does this code review work with GitLab and GitHub?▼

Yes. On GitLab it uses the glab:mr-status skill and its state script to dump the MR description, threads, and pipeline; on GitHub it uses gh pr view. Without an MR/PR, it falls back to master or main as the base branch.

How are false positive review findings filtered out?▼

Each finding is re-derived against the actual code and tested against six grounds: unreachable, already guarded, sanctioned convention, framework semantics misread, pre-existing, and impact inflated. Unconfirmed findings are dropped and the report's coverage section records the counts.

Can the review assess database migration safety?▼

Yes. When the diff contains a migration, the release review agent produces a safety assessment with a verdict of Run anytime, Run in low-traffic hours, or Do not run as written, citing engine version, lock behavior, table sizes, and documentation sources.

Where is the code review report saved?▼

The report is written to ./.claude/review-report/<topic>.md, where the topic is derived from the branch name or ticket ID. A brief inline summary with verdict, findings by severity, and coverage is also shown in the conversation.