{"slug":"protocol-claims","verification":{"valid":true,"entries":4,"head":"b501a13e39a1faeab6fbdb539da7c9be5c43fb2eb48dfc64f76e89864a73d6b3"},"energy":{"passes":4,"tokens_in":2718,"tokens_out":933,"tokens_total":3651,"cost_usd":0,"models":{"grok/build":1,"grok/grok-4.3":2,"owner":1},"head":"b501a13e39a1faeab6fbdb539da7c9be5c43fb2eb48dfc64f76e89864a73d6b3"},"provenance":[{"ts":"2026-06-29T16:53:49.095Z","model":"grok/build","action":"write","prompt":"post_protocol_meta_articles.mjs","input":"protocol-claims","response":"","tokens_in":0,"tokens_out":0,"cost":0,"prev":"genesis","hash":"42c03257a181a9871913d2cbbc8ad78edb9b00860c042796aa5d93151bb9917e"},{"ts":"2026-06-29T16:54:16.885Z","model":"grok/grok-4.3","action":"synthesize_body","prompt":"","input":"protocol-claims","response":"## Regeneration vs degeneration — where this fits\nThe body is always doing two things at once: breaking down and building back. A condition persists when breakdown outruns repair. Most drugs used for symptoms suppress a signal without fixing the tissue that caused the signal. This ledger maps compounds studied for repair pathways: new blood vessels, repair-cell migration, nerve regrowth, gut lining, neural connections.\n\n## What it is\nThis entry is a protocol for how claims are built and checked inside the ledger. It defines voxel atoms, tiers, weights, and honesty gates. No compound data is stored here. The entry tracks how any future claim gets counted or removed.\n\n## How it works\nA claim enters as a voxel atom.  \nStep 1: the claim is tagged with a tier (human, preclinical, anecdotal).  \nStep 2: a weight is assigned based on source type and quality.  \nStep 3: an honesty gate checks for retraction, cut, or challenge.  \nStep 4: active claims remain in the count; retracted or cut claims stay on the ledger but are excluded from active totals unless the query asks for inactive data.\n\n## Why it would work (logic chain)\nIf claims are not tagged and weighted, counts mix strong and weak data. If honesty gates do not run, retracted items stay mixed in. If tiers stay fixed, human data stays separate from animal data and from single reports. The system therefore produces counts that reflect source strength rather than volume alone.\n\n## Why people take it\nPeople do not take this protocol. It is a method for organizing data, not a substance.\n\n## How many people take it\nZero. The ledger records zero users for this entry.\n\n## Evidence inventory\nSources total: 0.  \nStudies catalogued: 0.  \nReddit posts: 0.  \nX posts: 0.  \nOther anecdote sources: 0.  \nClaims marked human: 0.  \nClaims marked preclinical: 0.  \nClaims marked anecdotal: 0.  \nNo human studies, no animal studies, no anecdotes are present.\n\n## What scientists say\nNo scientist statements are recorded for this entry.\n\n## Wha","tokens_in":1147,"tokens_out":521,"cost":0,"prev":"42c03257a181a9871913d2cbbc8ad78edb9b00860c042796aa5d93151bb9917e","hash":"9b729d7d57644012b48c7a7c0b72b4b52528aa59c805bc65636fd8f7ea0e07dc"},{"ts":"2026-06-29T17:00:59.660Z","model":"grok/grok-4.3","action":"synthesize_body","prompt":"","input":"protocol-claims","response":"## What this article is\n\nThis document explains how the miscsubjects.com claim system works. It defines voxel atoms, tiers, weights, and honesty gates that control what enters published articles.\n\nOne claim equals one falsifiable assertion. Compound claims violate the constitution and fail health audit. System tier is reserved for architecture and protocol axioms, not biological mechanism claims.\n\n## Who claims what\n\nSystem tier claims are authored by the protocol itself. They are not attributed to external studies or anonymous sources.\n\n## What is known\n\n- Posted_by.actor uses canonical format provider/model or system/function for server passes.\n- Source_ids must reference rows in the hash-chained source ledger. Orphan sources can be materialized into claims via repair.\n- Elevation threshold defaults to 0.6. Review contributions scoring below this do not enter the article.\n- Cut_threshold defaults to 0.15. Scored claims below this receive status:cut and drop from reader section order.\n- Adversary challenges append to meta.challenges[] and raise stance_scores.adversary on the target.\n- Retracted claims remain in meta.claims[] with status:retracted. Ask and bundle exclude them unless include_inactive=1.\n- Ingest creates evidence_ingest nodes on the question graph linking pasted text to new sources and claims.\n- Who_claims_what slot must name study authors, platforms, or n= — not anonymous \"studies show\".\n- Ledger_durability.js infers slot from section headers and retiers clinical_trial sources toward human when RCT language present.\n- Weight recomputation is monotonic per score pass. Reviews accumulate. Rest gate discards non-material passes as energy_spent.\n\n## What is unknown\n\nAutomated tier assignment from PubMed titles alone is incomplete. Abstract parsing remains a gap.\n\nCross-model claim deduplication is heuristic. Near-duplicate assertions may coexist until adversary or manual retract.\n\n## Limitations\n\nNo doses, no \"you should take\" — claims violating this fai","tokens_in":1571,"tokens_out":412,"cost":0,"prev":"9b729d7d57644012b48c7a7c0b72b4b52528aa59c805bc65636fd8f7ea0e07dc","hash":"ce726042fbe055a87e7cd50d1ba28a7d05249777a7df3e6c649d48a91bcdcf4b"},{"ts":"2026-07-17T02:41:02.013Z","model":"owner","action":"voxel_divide","prompt":"","input":"protocol-claims","response":"14 DIVs from body (verbatim, roundtrip-checked)","tokens_in":0,"tokens_out":0,"cost":0,"prev":"ce726042fbe055a87e7cd50d1ba28a7d05249777a7df3e6c649d48a91bcdcf4b","hash":"b501a13e39a1faeab6fbdb539da7c9be5c43fb2eb48dfc64f76e89864a73d6b3"}]}