Which backends are supported by flow-next-epic-review?

The skill currently supports RepoPrompt (rp) and Codex CLI (codex) as backends, with an option to skip the review gate by setting the backend to 'none'.

What is the difference between an epic review and an implementation review?

An epic review specifically verifies that the combined implementation satisfies the high-level specification requirements, whereas an implementation review typically focuses on granular code quality, style, and logic.

How do I trigger an epic review in Claude Code?

You can trigger the process using the command /flow-next:epic-review followed by the specific Epic ID, for example: /flow-next:epic-review fn-101.

Does this skill require manual intervention during the fix loop?

No, the skill is designed to autonomously parse issues from the reviewer feedback, apply fixes, run tests, commit changes, and re-review until a 'SHIP' verdict is reached.

What happens if a review backend is not configured?

The skill will look for a backend in arguments, environment variables, or the .flow/config.json file. If none are found, it will prompt the user to run the setup command.

Flow-Next Epic Review

Name: Flow-Next Epic Review
Author: gmickel

bygmickel

•

496

Security & Testing

Verifies that completed epic tasks align with original specification requirements using automated review gates and autonomous fix loops.

About

Flow-Next Epic Review is a specialized workflow coordinator designed to bridge the gap between feature implementation and specification compliance. Unlike standard code quality checks, this skill acts as an orchestrator that leverages external backends like RepoPrompt or Codex to ensure that the holistic output of an entire epic satisfies every requirement in the spec. It features a sophisticated internal fix loop that autonomously resolves reviewer feedback, committing changes and re-submitting for review until a definitive 'SHIP' verdict is achieved, significantly reducing the manual overhead of final feature validation.

Key Features

Configurable backend selection via environment variables, config files, or CLI flags
Autonomous internal fix-loop for resolving implementation gaps without user intervention
496 GitHub stars
Receipt-based session continuity for multi-turn reviews and re-anchoring
Automated spec-compliance verification via RepoPrompt or Codex backends
Mandatory commit-before-review logic to ensure version-controlled fix history

Use Cases

Finalizing complex multi-task features to ensure total alignment with original requirements
Automating the quality gate process in a plan-first AI coding workflow
Maintaining architectural integrity by preventing requirement drift over long development cycles

About

Key Features

Configurable backend selection via environment variables, config files, or CLI flags
Autonomous internal fix-loop for resolving implementation gaps without user intervention
496 GitHub stars
Receipt-based session continuity for multi-turn reviews and re-anchoring
Automated spec-compliance verification via RepoPrompt or Codex backends
Mandatory commit-before-review logic to ensure version-controlled fix history

Use Cases

Finalizing complex multi-task features to ensure total alignment with original requirements
Automating the quality gate process in a plan-first AI coding workflow
Maintaining architectural integrity by preventing requirement drift over long development cycles