Which government bodies do the stored document tables name, and how firm is the evidence behind each link?

Document links. Part of what the twin can answer. Every figure below is the value the route serves, with its source; the raw response is at /api/twin/brain/doclinks.

Descriptive, sourced: no model

Basis: sourced records (URL, sha256 or locator per row), no model.

As of: 24 September 2026.

Before, added and after

Evidence basis	Links	PDFs	Bodies	Share of added
Provenance same sha256	Links: 3,548	PDFs: 6	Bodies: 12

Match	Links	Bodies
Cell equals	Links: 5,174	Bodies: 32
Upazila column	Links: 351	Bodies:

Body	Bengali name	Level	Added links
Sylhet Division	Bengali name: সিলেট বিভাগ	Level: Division	Added links: 708
Khulna Division

Reason	Documents	Links

Question	Answer
How many body-to-document-table links does the twin serve now?	Answer: 14,768 links from 99 stored PDFs to 443 bodies (9,045 links from 52 PDFs to 408 bodies before this lane; 5,723 links from 54 PDFs added, 35 of them reaching bodies that had no link before).
How many of the links DOCLAYER2 withheld for a missing fetch URL are now served?	Answer: 8,350 links were withheld; 5,730 of them sit on 55 PDFs whose fetch URL was recovered from stored evidence, and 5,722 of those pass every gate and are served. 2,620 links on 115 PDFs stay withheld because no stored fetch record ties to the file.
How firm is the evidence behind the recovered URLs?

Source	Caveat
Body to document-table links (lane DOCLAYER2)	Caveat: A cell prints the body's exact registry name, or an upazila's name under an Upazila/Thana column with its district in the same table. A link says the body is named in the table, not that the table's figures are about it.
Links recovered from withheld PDFs, plus exact full-name field-office and union cells	Caveat: Same rules and gates as DOCLAYER2 (person-data tables excluded, PDF re-hashed equal to its recorded sha256, name found again in the stored page text). A fetch URL is recovered only from stored evidence; the row names its basis.
Per-PDF status of every PDF DOCLAYER2 or the extension would have linked

Source collection	Candidate links	Reason
Govtwin min agri	Candidate links: 1,190	Reason: No stored fetch record ties to this file
Govtwin sdg tracker	Candidate links: 185	Reason: No stored fetch record ties to this file
Dof bd	Candidate links: 87	Reason: No stored fetch record ties to this file
Dof bd extra	Candidate links: 87	Reason: No stored fetch record ties to this file
Govtwin procurement	Candidate links: 74	Reason: No stored fetch record ties to this file
Dof bd	Candidate links: 66	Reason: No stored fetch record ties to this file
Dof bd extra	Candidate links: 66	Reason: No stored fetch record ties to this file
Govtwin min fisheries	Candidate links: 66	Reason: No stored fetch record ties to this file
Govtwin min fisheries	Candidate links: 64	Reason: No stored fetch record ties to this file
CPD	Candidate links: 48	Reason: No stored fetch record ties to this file