iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
An AI code reviewer that cites your instruction file has shown you something, but not necessarily what you want to know. A citation with a rule name and a line range shows that the rule was surfaced to the reviewer. It does not show that the rule drove the finding, and it does not show that the cited text is still the current text. Daniel Nwaneri’s test of Qodo and CodeRabbit on a small Cloudflare Worker repository, published September 9, 2026 with a reader-driven correction on September 10, 2026, makes that gap concrete.
The question his post opens with is “Does your AI code reviewer read CLAUDE.md?” The more useful version is whether the reviewer can show which current line it read, and whether that line says what it claims.
What the test set up
The repository, rules-demo-api, is a small Cloudflare Worker that accepts feedback and logs a rating. It carries matching rules in CLAUDE.md and AGENTS.md, so the same instructions appear in both files. Two rules matter for the test:
Recommended Free Tools
- “Never log request headers or bodies. They may contain secrets, tokens, or PII.”
- “All new route paths must be kebab-case (
/feedback-summary), never camelCase (/feedbackSummary). This is an internal convention, not a general best practice.”
Each rule was tested by adding code that violates it, then reviewing the change with each tool.
#1 Best Overall
- Comprehensive Coverage: Dive deep into Python with thorough explanations of key topics and practical, real-world examples that make complex concepts easy to grasp. Our content is designed to provide you with a strong foundation and advanced skills, ensuring you are well-prepared for any Python-related challenge.
- Interactive Learning: Transform your learning experience with our interactive format. Practice and apply what you learn immediately with hands-on code snippets and exercises. This approach not only reinforces your understanding but also helps you develop practical coding skills that you can use in real projects.
- Portable Convenience: Take your learning journey anywhere with our highly portable resources. Whether you’re at home, on the commute, or traveling, you can study and code whenever it suits you. Our materials are accessible across devices, making it easy to fit learning into your busy schedule.
- Versatile Audience: Our content is tailored to meet the needs of a wide range of learners. Whether you’re a student looking to ace your exams, a professional aiming to advance your career, or a hobbyist passionate about coding, our resources are designed to help you achieve your goals.
- Skill Enhancement: Boost your confidence and retention with our regularly updated content. Stay ahead of the curve with the latest Python advancements and trends. Our continuously refreshed materials ensure that you are always learning the most current and relevant information, keeping your skills sharp and up-to-date.
Scenario one: a risk visible without the file
The first change added logging of request headers and body. Both Qodo and CodeRabbit flagged it. The author treats this as a weak test of file-reading: logging credentials-bearing request data is recognizable as risky without any repository instruction, so a catch here tells you little about whether the local rule was consulted.
Scenario two: the kebab-case route
The second change added a camelCase route, /feedbackSummary, which the kebab-case rule forbids. Results differed by tool and by profile.
Qodo
Qodo cited the kebab-case rule and reasoned about a route mismatch. This is the closest the test came to a visible link between a finding and a repository instruction.
Rank #2
- Comprehensive Coverage: SQL Flashcards and NoSQL Flashcards designed for beginners and interview prep, covering core database concepts, queries, indexing, normalization, and real-world use cases. From relational structures, JOINs, and indexing to NoSQL document models, key-value stores, and distributed systems, these flashcards give you a solid foundation and advanced knowledge to handle any database challenge confidently.
- Interactive Learning: Enhance your understanding with an interactive, hands-on approach. Each card includes practical query examples, schema illustrations, and exercises that let you immediately apply what you learn. This active learning style helps you strengthen your querying skills and build intuition for solving real data problems. Beginner-friendly explanations that help you learn SQL and NoSQL faster without overwhelming theory or dense textbooks
- Portable Convenience: Study databases anytime, anywhere. Whether you’re at home, commuting, or taking a break, these portable flashcards make it easy to learn on the go. Perfect for busy students, developers, or professionals fitting learning into a tight schedule.
- Versatile Audience: Designed for all learners from students preparing for exams to data analysts, backend engineers, and tech enthusiasts. Whether you're building your first query or optimizing production databases, these flashcards guide you at every stage of your learning journey. Perfect for SQL interview preparation for software engineers, data analysts, backend developers, and computer science students
- Skill Enhancement: Boost your confidence and stay current with evolving database technologies. Ideal for self-study, bootcamps, university courses, and last-minute interview revision with concise, memorable flashcard format
CodeRabbit on the Chill profile
CodeRabbit’s Chill profile returned no actionable comments on the route.
CodeRabbit on the Assertive profile
After switching the profile to Assertive, CodeRabbit flagged the naming issue. It did not refer to either instruction file.
The author also concedes that kebab-case is a common convention beyond this repository. A correct naming finding is therefore consistent with the reviewer applying general knowledge, and the result does not establish that CodeRabbit read the local rule.
Rank #3
- Comprehensive Coverage: Dive deep into JavaScript with thorough explanations of key topics and practical, real-world examples that make complex concepts easy to grasp. Our content is designed to provide you with a strong foundation and advanced skills, ensuring you are well-prepared for any JavaScript-related challenge.
- Interactive Learning: Transform your learning experience with our interactive format. Practice and apply what you learn immediately with hands-on code snippets and exercises. This approach not only reinforces your understanding but also helps you develop practical coding skills that you can use in real projects.
- Portable Convenience: Take your learning journey anywhere with our highly portable resources. Whether you’re at home, on the commute, or traveling, you can study whenever it suits you, making it easy to fit learning into your busy schedule.
- Versatile Audience: Our content is tailored to meet the needs of a wide range of learners. Whether you’re a student looking to ace your exams, a professional aiming to advance your career, or a hobbyist passionate about coding, our resources are designed to help you achieve your goals.
- QR Code Embedded: A QR code is embedded on each card at the top. At any point, if you need further clarification on a topic, simply scan the QR code with your smartphone. The QR code will take you to a YouTube video or an article that provides a detailed explanation of the topic.
The reader’s objection
A reader, pm25coder, made the central methodological point in the discussion: a citation proves a rule was surfaced, not that the file was read at decision time. Because the tested rule appeared in both CLAUDE.md and AGENTS.md, the experiment also could not show which file took precedence. Any citation to either file was consistent with either one having been used, or with the reviewer drawing on the same content from elsewhere.
Moving the rule, and what the citation still said
The author moved the kebab-case rule seven lines down in both files and reran Qodo. The finding still cited the old line ranges. The discussion draws two conclusions from this:
- The post-move review had a current commit anchor, but its prose label was stale. The reference looked up-to-date at the commit level while pointing at text that no longer sat at those lines.
- CodeRabbit had not rerun against the later commit, so its earlier output could not be compared with the moved rule.
The output alone does not reveal whether the reviewer was reading the current file, a cached copy, or an older version of the rule. A line range in a finding is a claim about the text, and it needs to be checked against the revision the review ran on.
What a citation can and cannot establish
The author’s own summary of the lesson is blunt: “A citation proves a rule was surfaced, not that it drove the decision or that it’s still accurate.” The discussion condenses it further: “So: citation proves surfaced, not causal, not current.”
In practice, a citation earns trust only after three checks:
- Attribution: the rule is unique to one file, so the finding can be tied to that file.
- Revision: the reference resolves to the exact commit that was reviewed.
- Content: the cited span actually contains the rule the finding names.
A catch that also matches general practice satisfies none of these by itself. The author’s closing observation on the two tools captures the practical difference: “Both tools caught the second one eventually. Only one of them could tell me why.” Even that “why” was only as good as the reference behind it.
Best Value
- [THE VIRAL 2026 TREND] Whether they are a "tech wizard" or just a fan of internet culture, this red lobster is the iconic symbol of 2026 success. Don't give a boring, generic card—give the one that shows you’re tuned into the latest trends and memes of their graduation year!
- [PROUD PARENT'S SECRET WEAPON] Want to be the "cool mom" or "cool dad"? This card is the perfect way to show your son or daughter that you truly "get" their world. Even if you don't know the code, they'll be impressed that you found the "Your Lobster is Ready" meme!
- [FOR EVERY 2026 GRADUATE] While it's a "must-have" for STEM majors, its quirky charm appeals to any grad who spent years "grinding." It’s the ultimate 'Let them cook' card—signaling that their hard work is finally complete and they are ready to deploy into the real world!
- [PREMIUM QUALITY & KEEPSAKE] Printed on 300gsm heavy-duty premium cardstock. It’s thick, durable, and perfect for displaying on a dorm room desk or office shelf as a souvenir of the year AI changed everything.
- [BLANK INSIDE FOR PERSONAL PROMPTS] The witty front sets the stage, leaving the inside blank for your heartfelt advice, funny memories, or a "bug-free" future wish. Includes a high-quality envelope, ready for immediate gifting.
A test that could attribute a finding
The author and commenters have proposed a stronger design. It has not been reported as run, so treat it as a method rather than a result:
- Place a unique sentinel string in
CLAUDE.mdonly. Choose a string that appears nowhere else in the codebase, so any citation of it must come from that file. - Place a conflicting rule in
AGENTS.mdonly, covering the same behavior with a different requirement. A citation then reveals which file was consulted. - Commit the change, record the commit SHA, and run each reviewer against that commit.
- Resolve every cited path and line range at that SHA, then confirm the span contains the named rule text.
- Rerun after each edit to a rule file, and record the new SHA with each result. A rerun on an older commit does not test the current text.
Where a rule can be expressed as a machine-checkable requirement, prefer that form. A required literal or a testable route pattern, such as ^/[a-z0-9]+(-[a-z0-9]+)*$ for kebab-case paths, can be verified mechanically. A prose style instruction can only be judged by reading it, which is where the ambiguity in this test came from.
Comparison of the reported results
| Check | Qodo | CodeRabbit (Chill) | CodeRabbit (Assertive) |
|---|---|---|---|
| Flagged logging of headers and body | Yes (reported) | Yes (reported; profile for this scenario not stated) | Not stated |
| Flagged camelCase route | Yes, and cited the kebab-case rule | No actionable comments | Yes, without referring to either instruction file |
| Cited the rule by name and file | Yes (reported) | Not applicable | No |
| Rule unique to one file | No, duplicated in both files | No, duplicated in both files | No, duplicated in both files |
| Reference matched the moved rule after the seven-line move | No, still cited the old ranges | Not stated; not rerun on the later commit | Not stated; not rerun on the later commit |
| Reran on the changed commit | Yes (reported) | No (reported) | No (reported) |
The table reflects one author’s account of one repository. It is not a benchmark, and it says nothing about how either product performs across other codebases.
Limits of this account
The evidence is a single first-person write-up plus its discussion and the author’s reply. The test involved two tools, one repository and a handful of changes. It was not independently replicated, and this account does not verify current Qodo or CodeRabbit documentation, default settings, or profile behavior. Both products change, so a result from September 2026 is a snapshot of that period.
What to do with a citation
Treat a cited rule name and line range as a pointer to check, not as proof of reading. If a reviewer’s finding depends on a repository rule, confirm that the rule is unique to one file, that the reference resolves at the reviewed commit, and that the span says what the finding claims. A catch that matches general practice is a good catch, but it is not evidence that the local instruction file was used.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
