k-fin-full-review

name	k-fin-full-review
description	Use when the user asks for /full-review, a full k-fin agent-team review, a parallel security/platform/code/architecture review, or the Claude Code full-review equivalent in Codex.

k-fin Full Review

Run the Codex equivalent of .claude/commands/full-review.md.

Source files

• Command: .claude/commands/full-review.md
• Agent registry: AGENTS.md
• Agent definitions: .claude/agents/*.md

Workflow

1. Read AGENTS.md and .claude/commands/full-review.md.
2. Treat .claude/commands/full-review.md as the source of truth for reviewer set, review scope, and unified report shape.
3. Use AGENTS.md only to map each requested Claude agent to the closest Codex role.
4. Treat /full-review as explicit authorization to use parallel Codex subagents.
5. Ask each subagent to read its source Markdown before analyzing the repo.
6. Merge the returned reports according to .claude/commands/full-review.md.

Do not invent findings. If a reviewer reports no issues, keep that signal in the summary instead of padding the report.