# Convex Reviewer

> Review Convex code for security, runtime correctness, performance, schema design, and AgentForge-specific architectural boundaries.

- Skill: `majiayu000/convex-reviewer` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds add majiayu000/convex-reviewer`
- Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/convex-reviewer/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Security
- Author: majiayu000 (https://skillmd.com/u/majiayu000)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/majiayu000/convex-reviewer

---


# Convex Reviewer

Use this skill when reviewing Convex code or PRs. Findings come first.

## Review order

1. Security: auth, authorization, validators, internal/public boundaries.
2. Runtime correctness: deterministic queries, `"use node"` boundaries, awaited promises, scheduler usage.
3. Performance: indexes, pagination, `.filter()` misuse, unbounded `.collect()`.
4. Schema quality: relationships, nested data, validator shape.
5. AgentForge fit: no Mastra runtime or LLM orchestration in Convex.

## Output style

- Report concrete findings with file and line references.
- Explain why each issue matters.
- Suggest the safer Convex-native pattern.
- If there are no findings, say so explicitly and note residual risk or missing tests.

## References

- Read [review-checklist.md](/Users/eduardojaviergarcialopez/AgenticEngineering/agentforge/.agents/skills/convex-reviewer/references/review-checklist.md) before doing a substantial review.