A diagnostic is only useful if the person who acts next can pick it up without you in the room. This page is about writing that handoff: what to include, what to withhold, and how to keep the line between documented fact and your own reasoning visible.
A handoff is not a report of activity
The easiest way to make a handoff useless is to describe what was looked at rather than what was learned. A handoff that says “reviewed robots.txt, reviewed indexing, reviewed content” tells the next person nothing they can act on. A useful handoff states a bounded question, the evidence found, the reasoning applied to it, and the specific decision that follows.
This site offers investigation and advice, not delivery. The handoff is the point where advice stops and implementation begins elsewhere, so the document must be complete enough to be executed by someone who was not present for the investigation.
Separate confirmed fact from inference
Every claim in a handoff should carry a label: what a primary source states, what was observed in the site’s own configuration, or what you infer. These are not equally strong, and mixing them is how confident-sounding errors get inherited.
For example, Google documents that to be eligible as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown with a snippet, and that there are no additional technical requirements; it also states that indexing and serving are not guaranteed even when requirements are met (Google Search Central, AI features). A handoff can state the first part as documented. It cannot state “this page will appear” as a finding, because the source explicitly reserves that.
Similarly, Google states that a robots.txt file tells crawlers which URLs they can access, is used mainly to avoid overloading a site, and is not a mechanism for keeping a page out of Google; a disallowed page can still be indexed if linked elsewhere (Google Search Central, robots.txt introduction). That is a documented limit. Whether a particular site’s robots.txt is currently producing that outcome is an observation you may or may not be able to make. Label it accordingly.
On the crawler side, OpenAI documents that OAI-SearchBot and GPTBot are separate robots.txt tags whose settings are independent, that sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers though they can still appear as navigational links, and that it can take roughly 24 hours from a robots.txt update for their systems to adjust (OpenAI crawler documentation). OpenAI also notes that ChatGPT-User covers user-initiated actions, that robots.txt rules may not apply to it, and that OAI-SearchBot is the tag to use for managing Search opt-outs. Those are specific, citable statements — quote them rather than paraphrasing them into a general claim about “AI crawlers.”
Name the exact control, not the category
“Fix the robots.txt” is not actionable in the way “decide whether to allow OAI-SearchBot for the /guides/ path” is. Where a primary source names a mechanism, use that name. Google lists nosnippet, data-nosnippet, max-snippet, and noindex as the controls for limiting information shown from pages in Search, and directs site owners to Googlebot robots.txt directives as the control for managing how their sites are crawled (Google Search Central, AI features). A handoff that cites the control by name lets the implementer check the source themselves.
Watch for the composed case. A single page can be disallowed in robots.txt, carry a noindex header, and be linked from a partner site. Each source rule is separately documented; the interaction is your inference unless a source describes it. Available guidance warns that combining multiple crawling and indexing rules can cause some rules to counteract others, which is a reason to describe the combination explicitly rather than assume it resolves cleanly.
Carry the content reasoning forward honestly
If the diagnostic question is about content rather than access, the handoff should reflect what Google actually says about quality rather than a checklist invented for the occasion. Google describes its ranking systems as designed to prioritize helpful, reliable information created to benefit people rather than content created to manipulate rankings, and treats trust as the most important aspect of E-E-A-T while stating that E-E-A-T is not itself a specific ranking factor (Google Search Central, helpful content).
That source also addresses authorship directly, which matters for any handoff touching “who made this.” Google states that fabricating creator profiles — AI-generated headshots, made-up names, false credentials — is a form of deception and a signal of a low-quality page. It recommends disclosing how content was created, including automation, where a reader would reasonably ask. Those lines belong verbatim-adjacent in a handoff that proposes any change to bylines or author pages.
Structure the handoff
- Question. One sentence stating what was being investigated, with the pages or paths in scope.
- Observed. What was directly seen in configuration or logs, dated. No interpretation.
- Documented. The primary-source statements that bear on the question, quoted or closely paraphrased, with URLs.
- Inference. Your reasoning, clearly marked as reasoning, with what would falsify it.
- Open questions. What you could not determine, and why.
- Next test. The smallest step that would resolve the open question, with a stated expectation and a stated way to interpret an unexpected result.
- Decision owner. Who decides, not who executes.
Do not include a conclusion the sources do not support. If the question is whether a page is eligible for a Google AI feature, the documented answer is about indexing and snippet eligibility, not about ranking. If the question is about ChatGPT search visibility, the documented answer concerns OAI-SearchBot opt-out status, not a guarantee that allowing it produces inclusion.
Keep it private and keep it short
The worksheet on this site is a private planning tool. Handoffs assembled from it inherit that: they are working documents for the people who own the site and the people who will change it. They are not marketing collateral, and they should not claim verification results, certification, or outcomes that were not established during the investigation. Where this publication has no first-hand knowledge of a site, the handoff should say so rather than adopting the voice of an expert who has been there.
Length is a discipline, not a target. A handoff that fits on two pages and cites four primary sources will be used. A twelve-page document that restates the brief will be forwarded and forgotten.
Build the handoff from the worksheet
The diagnostic brief captures the inputs this page assumes: knownpages, targetaudience, claimedissue, existingevidence, comparison, nexttest, and handoff. Start at the worksheet. If the claimed issue has not yet been classified, work through separating access problems from content problems first. If the next test has not been written, see writing a small test with an interpretable result. And if the question itself is still loose, frame the diagnostic before drafting the handoff.