A site we do not own · published redacted · with permission
The second sample is somebody else’s site, and we cannot tell you whose.
The first sample is the operator’s own site. This one is a 25-page marketing services site, audited under the same rubric on the same terms, published with its owner’s permission and its identity removed. It is called TLN throughout.
- Score
- 57.1%44.5 of 78 resolvable points
- Grade
- CMaterial discoverability friction
- Coverage
- 78%9 checks inconclusive, 22 points unscored
- Pages read
- 2548 requests, 63 seconds
- Evidence
- 42hashed records, held by the operator
- Backlog
- 27items, ranked impact against effort
What was removed
Removed from every published file: the host name, the organisation name, four staff pages whose URLs are the people’s names, the city, and a named information-security certification. Each could identify the site.
No rating, observation, score, denominator or reason was changed.
- The hashes do not verify. Each finding carries the content hash of evidence captured from the original unredacted pages, so you cannot recompute them from this document. They show only that the operator holds a specific artifact.
- The JSON appendix is not published. It is 656 KB of captured page structure and observations in the site’s own words, and hundreds of verbatim phrases would still identify it after removing names.
- The redaction is a script, not a manual edit. It writes nothing if a single identifying term survives, and the published files are checked against it on every push.
If those limits matter, read the unredacted sample. It is complete, has its appendix, and its hashes verify.
Two of the four artifacts
-
The redacted report
All 33 checks, their ratings, the operator’s reason for each assisted one, the limitations, the page table and the evidence index. Nine checks are inconclusive.
report.html → -
The redacted remediation prompt
The paste-ready brief for whoever does the fixing, built only from findings that carry evidence. It proposes no fix for anything the audit could not establish.
fix-prompt.md → -
The JSON appendix and the operator’s decisions
Withheld. Both contain too much of the site’s own wording to redact. The same files for the unredacted sample, from the same code, are at /sample/.
What this sample shows
- Nine of thirty-three checks are inconclusive here, against three in the first sample, because the operator could answer his own intake questions. Those 22 points leave the denominator instead of scoring zero, so the score is 57.1 % of the 78 points this run could resolve.
- Topical coverage scores 34.1 and citation readiness 43.8, both F.
- Structured data scores 95.
- No percentile is quoted; the corpus is not yet large enough.