Browse categories
Explore
Fiverr Pro
English
$
USD
I will audit and fix your ai coding agent rules, skills, hooks and mcp setup
Max
About this gig
I will audit and restructure your AI coding-agent setup so your rules load reliably and critical constraints are enforced instead of buried in long markdown.
I work with Codex CLI, Claude Code, Cursor-style setups, AGENTS.md, CLAUDE.md, SKILL.md files, hooks, MCP servers, and command policies.
You will get:
- Root-cause diagnosis of skipped rules
- Load-path and truncation checks
- Clean replacement AGENTS.md / CLAUDE.md structure
- Rewritten SKILL.md files with stronger trigger descriptions
- Hook / exec-policy recommendations for machine-checkable rules
- MCP pruning suggestions to reduce context bloat
- Verification steps so you can confirm the setup works
I do not rewrite your app or fix unrelated code bugs. This service is focused on agent configuration, instruction architecture, rule reliability, and enforcement design.
Best for teams or developers whose AI coding agent ignores project rules, skips skills, touches protected files, or forgets required workflows.
Get to know Max
Max
- FromUnited Arab Emirates
- Member sinceFeb 2023
- Avg. response time2 hours
- Last delivery2 years
Languages
Arabic, English
I help founders and businesses turn ideas into reliable AI-powered software, automation, and production-ready systems.
I work with Claude Code, Cursor, Codex, Lovable, MCP, AI agents, Python, JavaScript, APIs, and modern automation tools.
My focus is on clean execution, smart architecture, fast problem-solving, and dependable delivery—whether you need an MVP, a critical fix, an AI workflow, or a stronger existing product.
FAQ
Q: Can you guarantee the AI agent will never skip rules again?
A: No AI setup can guarantee perfect obedience, but I can reduce failures by shrinking always-loaded instructions, improving trigger design, and moving machine-checkable rules into hooks or command policies.
Q: Do you fix code bugs too?
A: No. This gig is for coding-agent configuration, rules, skills, hooks, MCP setup, and instruction architecture.
Q: Do you need my full repository?
A: Usually no. I need instruction files, skill/rule files, config excerpts, MCP list, and examples of skipped rules. A sanitized repo is fine.
Q: Can you work with both Codex and Claude Code?
A: Yes. I can review Codex-style AGENTS.md setups and Claude Code CLAUDE.md / rules / hooks setups.
Q: What do I receive?
A: Depending on the package: diagnosis table, replacement files, hook recommendations, MCP prune list, migration checklist, and verification prompts.
