What the saved FOIA records show

Question 1

What is in this FOIA archive?

Accepted Answer

This preserved export contains 195 request threads, 1,824 saved messages and 1,385 attachment entries, grouped under 62 agency labels after case-merging. It covers October 17, 2024 through February 25, 2025. These are counts of the saved export, not a complete census of all later requests, responses, agencies or findings.

Question 2

How do I check a connection?

Accepted Answer

Open the connection map, search a label and follow its exact FOIA communication link. Each source anchor identifies one numbered message and its exported date. Read the full message and the quoted material around the match. A repeated case number, email address or legal citation can be copied into many threads and is not independent corroboration.

Question 3

Do the dots prove employment, coordination or wrongdoing?

Accepted Answer

No. Request-to-agency edges come from saved recipient metadata. Other edges come from text or signature-pattern matching. A name, email address, quoted signature or repeated paragraph does not authenticate authorship, prove a current role, establish employment or coordination, or demonstrate misconduct. The map supplies source links for review, not legal findings.

Question 4

Do five threads prove that an agency never replied?

Accepted Answer

No. The original heuristic flags 5 threads without a message it classified as an agency reply. Its sender test can misread signatures and quoted letters. The export also stops in February 2025. This does not prove no reply was sent, no later reply exists, or any response deadline was violated.

Question 5

Were records actually released in seven requests?

Accepted Answer

The phrase matcher labeled 7 threads “records released.” Its expression also matches “final response,” which may not contain released records. This is a reproducible text-pattern count, not a verified count of successful releases. Read the linked response and attachments to determine what was actually supplied.

Question 6

Can I open every saved attachment?

Accepted Answer

The request pages provide the saved attachment entries and the 195 agency.txt recipient-source files. An attachment may be referenced in more than one thread; entry counts are not unique-file counts. One oversized original PDF is offered inside a lossless ZIP wrapper. Extracting that ZIP restores the original PDF bytes. The original attachment bytes are preserved and full SHA-256 fingerprints allow byte-identity checks; a matching hash does not establish the truth of the contents.

Question 7

Were any public message details masked?

Accepted Answer

The public message projections mask 25 portal-credential assignments. This protects access credentials; it is not a finding about the agency or the request. The preserved originals remain unchanged in the local source vault. This exception applies to the message projections, not recompression or rewriting of the original attachment files.

Question 8

Is this the same inventory as the older MuckRock request list?

Accepted Answer

The older published inventory and this preserved 195-request export are distinct source snapshots. Their scopes and dates must be kept separate. The new export adds complete saved message threads and attachment access for this snapshot; it does not silently replace earlier source records or establish what happened after February 25, 2025.

Question 9

Does publication mean Google, Bing or an AI has indexed every record?

Accepted Answer

No. A working page, canonical URL, sitemap entry or readable structured record establishes technical availability. Submission, crawling, indexing and an observed AI citation are separate events that need their own dated evidence. No ranking, indexing or universal AI-discovery result is promised.

What the saved FOIA records show — and what they do not