Browse categories
Explore
Fiverr Pro
English
$
USD
I will audit your ai coding agent for reliability and context waste
David Wieczorek
About this gig
Is Codex, Claude Code, or Cursor ignoring instructions, rereading files, looping, or struggling to verify changes? I will audit your AI coding agent and repository to find the reliability bottlenecks behind wasted context and inconsistent output.
You receive an evidence-backed report covering:
conflicting or duplicated instructions
context waste and repeated work
tool, memory, and workflow risks
missing tests and weak verification
high-impact fixes in priority order
Basic: repo scan + prioritized findings.
Standard: full instruction/context/tool/test audit.
Premium: full audit plus bounded fixes, regression tests, and verification.
Send a repository or project export, the agent instructions, the problem you are seeing, and your expected behavior. Never send passwords, API keys, private tokens, or secrets. I work with Codex, Claude Code, Cursor, and similar AI coding workflows. Scope is one repository or project per package; large multi-repo migrations require a custom offer.
Get to know David Wieczorek
David Wieczorek
Freelancer
- FromSweden
- Member sinceJun 2020
- Avg. response time1 hour
Languages
Swedish, English
I help founders and small teams diagnose unreliable AI-built software. I work with Codex, Claude Code, Cursor, Lovable, Bolt, Replit, and similar tools. I audit repositories and applications for conflicting instructions, context waste, repeated work, missing tests, weak verification, broken workflows, and production-readiness risks. You receive a clear, evidence-backed report explaining what is wrong, why it matters, and what to fix first. I separate measured facts from recommendations and unknowns. I can also implement bounded fixes, create reliability tests, and verify the results.
Other AI Development Services I Offer
FAQ
What do you need from me to start?
Send a repository or project export, the relevant agent instructions, the problem you are seeing, and your expected behavior. Please remove passwords, API keys, tokens, and other secrets.
Can you work with Codex, Claude Code, or Cursor?
Yes. I audit the repository, instruction files, tool workflow, context usage, tests, and verification behavior for these and similar AI coding workflows.
Will you fix the issues too?
The Premium package includes bounded fixes, regression tests, and verification. Large rewrites, new features, or multi-repository work can be quoted separately.
Do you need account access or credentials?
No. I do not need passwords, API keys, private tokens, or direct account access. Share a sanitized repository or export and relevant instructions instead.
What will I receive?
You will receive a clear report with evidence, risk explanations, prioritized recommendations, and next steps. Where included, you also receive the bounded changes and verification results.
