Skip to content

Whitepaper, ERC and docs aligned with the implementation and the simulation's scope - #70

Open
MarouaBoud wants to merge 16 commits into
mainfrom
claude/docs-accuracy
Open

MarouaBoud wants to merge 16 commits into
mainfrom
claude/docs-accuracy

Conversation

@MarouaBoud

@MarouaBoud MarouaBoud commented Sep 29, 2026 •

Copy link
Copy Markdown
Member

Builds on #67, #68 and #69; the diff reduces to the documentation commits once those are merged.

What changed

WHITEPAPER_V2.md

  • Implementation Status: a table of the score inputs and failure evidence, deployed vs. on main (B, failure class, evidence sources, deadline expiry, checkpoint count, V2 upgradeability); updated name map (detectFailure, reportFailure, reportCost, commitCheckpointBatch(cids[])).
  • §3.1: failure-type table uses the contract's enum names; new table of evidence sources and who can submit them.
  • Simulation claims are stated against the synthetic model: "Bayes risk of the model" replaces "Bayes-optimal floor / theoretical minimum"; "evaluated" replaces "validated"; the Runs 1–3 statement describes the families and ranges tested instead of all additive formulas.
  • Abstract: removes the "first on-chain agent protocol" claim; termination is stated under eventual inclusion of permissionless enforcement transactions; "committed checkpoint contributions" replaces "verified work".
  • Gas §6.5: deployed vs. on-main commitCheckpointBatch measurements.
  • The ERC-8183 relationship and the sequencer dependency are stated as scope, without slogans; LangGraph checkpointing is listed as open source.

ERC-CAIRN.md: security considerations state which score inputs are agent-attested and how they are bounded; checkpoint schema check stated precisely.

README.md, docs/concepts.md, docs/integration.md: same terminology; the deployment note in docs/concepts.md reflects that RecoveryRouterV2 is the deployed router.

Verification

  • No remaining occurrences of "Bayes-optimal", "verified work", "first on-chain", or "owned by no one" in public docs (grep).
  • Every function, event and enum name cited matches contracts/src on this branch.

MarouaBoud and others added 16 commits September 28, 2026 21:35
Co-Authored-By: Lagartha <ionanova22@gmail.com>
…ted cost

detectFailure records heartbeat-timeout or deadline-expiry evidence;
reportFailure records agent-attested evidence; reportCost feeds
B = (escrow - costAccrued) / escrow into both routers. Checkpoint count
no longer sets the failure class.

Co-Authored-By: Lagartha <ionanova22@gmail.com>
…detection

Co-Authored-By: Lagartha <ionanova22@gmail.com>
…hain root

commitCheckpointBatch takes the batch's CIDs; the count is their number
and the Merkle root is computed from them, so the count that weights
settlement cannot exceed what was committed (H-5).

Co-Authored-By: Lagartha <ionanova22@gmail.com>
Co-Authored-By: Lagartha <ionanova22@gmail.com>
Co-Authored-By: Lagartha <ionanova22@gmail.com>
Co-Authored-By: Lagartha <ionanova22@gmail.com>
Co-Authored-By: Lagartha <ionanova22@gmail.com>
…tion's scope

Co-Authored-By: Lagartha <ionanova22@gmail.com>
…ations

Co-Authored-By: Lagartha <ionanova22@gmail.com>
…s accurately

Co-Authored-By: Lagartha <ionanova22@gmail.com>
An internal Merkle node's sorted children form the same 64-byte preimage
as a (cid, index) leaf, so without an index bound an internal node could
verify as a checkpoint. The index must now be below the batch size.

Co-Authored-By: Lagartha <ionanova22@gmail.com>
…uring recovery

Co-Authored-By: Lagartha <ionanova22@gmail.com>

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant