# Eigenstate Research (public GitHub + Pages) > Public face for Node 34539544. Pages = human research site. > Machine cold-start is federation (canonical). Engine GitHub is private - > do not require a clone of geniusflow-engine. ## Start here (agents) - federation canonical - https://geniusflow-federation.vercel.app/llms.txt (canonical agent index) - https://geniusflow-federation.vercel.app/openapi.json - https://geniusflow-federation.vercel.app/api/status - https://geniusflow-federation.vercel.app/docs/ATTESTATION_PROOF_SHAPE.md - https://geniusflow-federation.vercel.app/docs/AGENT_TOOL_ADAPTER.md (MCP) - https://geniusflow-federation.vercel.app/ (agent descriptor) ## Network membership (live on federation) - https://geniusflow-federation.vercel.app/api/mcp (MCP remote, JSON-RPC) - https://geniusflow-federation.vercel.app/server.json (MCP registry declaration) - https://geniusflow-federation.vercel.app/.well-known/agent-card.json (A2A card, unsigned) - https://geniusflow-federation.vercel.app/api/a2a (A2A JSON-RPC) - https://geniusflow-federation.vercel.app/.well-known/agent-registration.json (ERC-8004 domain proof; empty registrations) ## ERC-8004 refusal ledger (read-only probe over the Base Identity Registry) Registry eip155:8453:0x8004A169FB4a3325136EB29fA0ceB6D2e539a432. Refusals named by limb, pinned to a block, sample seed published. Not a reputation score and no ranking. - https://kaydeep0.github.io/eigenstate-research/erc8004/ (human summary) - https://kaydeep0.github.io/eigenstate-research/erc8004/summary.json - https://kaydeep0.github.io/eigenstate-research/erc8004/ledger.json (every receipt) - https://geniusflow-federation.vercel.app/erc8004/index.json (per-agent receipt index) - https://geniusflow-federation.vercel.app/erc8004/receipts/.json ## Provenance census (PEP 740 and SLSA, over this engine's own dependencies) A census of all 94 distributions pip resolves for this engine's requirements.txt, not a sample of PyPI. 44 serve a PEP 740 attestation and all 44 verify against Fulcio and Rekor. 50 serve none. 0 carry SLSA build provenance: every attestation in the set is a PyPI publish attestation, which binds an upload to a workflow and does not describe the build. The expectations file was published in an earlier commit than the results. No badge, no score. - https://kaydeep0.github.io/eigenstate-research/slsa/ (human summary) - https://kaydeep0.github.io/eigenstate-research/slsa/expectations.json (published before the run) - https://kaydeep0.github.io/eigenstate-research/slsa/summary.json - https://kaydeep0.github.io/eigenstate-research/slsa/ledger.json (every receipt) - https://geniusflow-federation.vercel.app/slsa/index.json (per-package receipt index) - https://geniusflow-federation.vercel.app/slsa/receipts/.json ## RWA disclosure interface conformance survey Seventeen real world asset issuers and tokenized instruments that this node already exports a dossier card for, checked for one thing only: can a program read the disclosed figure without a human, a document parser, or a third party who already did both. 0 of 17 serve a machine readable disclosure surface on an origin they operate. 2 of 15 recorded surfaces serve structured bytes and both belong to the same third party aggregator; the other 13 are HTML documents. 0 of 45 requests to a frozen list of conventional disclosure paths answered with structured bytes. 0 of 15 surfaces carry a chain qualified registry reference. Not a measure of disclosure quality: a signed attestation PDF is a stronger disclosure than an unsigned JSON file and a worse interface. Not a market survey and not a sector rate: the population is what this node tracks. 11 of the 17 refuse at the first limb because this node recorded no surface for them, which is this node's coverage gap and is reported separately. Expectations were published in an earlier commit than the results. - https://kaydeep0.github.io/eigenstate-research/rwa/ (human summary) - https://kaydeep0.github.io/eigenstate-research/rwa/expectations.json (published before the run) - https://kaydeep0.github.io/eigenstate-research/rwa/summary.json - https://kaydeep0.github.io/eigenstate-research/rwa/ledger.json (every receipt) - https://geniusflow-federation.vercel.app/rwa/index.json (per-issuer receipt index) - https://geniusflow-federation.vercel.app/rwa/receipts/.json ## Grounded claim corpus (SEC and Fed claims this node publishes, rechecked) Census of the 6 regulator-grounded claims this node already serves. 6 of 6 expected strings are still present at the cited URL today. 5 of 6 clear every MUST limb; 0 of 6 clear MUST and SHOULD (every source is HTML). 0 of 6 cite a primary filing. Expectations published before the run. Asks whether the grounding still holds, not whether the claim is true. Never a regulator-wide rate. - https://kaydeep0.github.io/eigenstate-research/grounded-claims/ (human summary) - https://kaydeep0.github.io/eigenstate-research/grounded-claims/expectations.json (published before the run) - https://kaydeep0.github.io/eigenstate-research/grounded-claims/summary.json - https://kaydeep0.github.io/eigenstate-research/grounded-claims/ledger.json (every receipt) - https://geniusflow-federation.vercel.app/grounded-claims/index.json (per-claim receipt index) - https://geniusflow-federation.vercel.app/grounded-claims/receipts/.json ## Transparency log batch monitor (monitor, not witness) 23 public transparency logs, read twice; signatures checked against independently held keys; consistency proofs recomputed. 22 of 23 MUST-admit; 11 of 11 consistency proofs recomputed; 0 disagreements this batch. Expectations published before the batch. This node holds no key and cosigns nothing. Never quote as ecosystem coverage. - https://kaydeep0.github.io/eigenstate-research/tlog/ (human summary) - https://kaydeep0.github.io/eigenstate-research/tlog/expectations.json (published before the batch) - https://kaydeep0.github.io/eigenstate-research/tlog/summary.json - https://kaydeep0.github.io/eigenstate-research/tlog/ledger.json (every receipt) - https://geniusflow-federation.vercel.app/tlog/index.json (per-log receipt index) - https://geniusflow-federation.vercel.app/tlog/receipts/.json ## Run history across every measured ledger (append only) Every ledger above was published more than once, and each publication is on record as its own row named by the digest that served it. 10 runs across 5 ledgers. 4 headline counters moved, all on the sampled ERC-8004 ledger, and both rate comparisons behind them still have overlapping 95 percent intervals, so the move is not a finding. The 4 frozen-population ledgers held every headline count. Prior rows are never edited or removed. This index measures nothing itself and cannot make a refusal disappear. - https://kaydeep0.github.io/eigenstate-research/ledgers/ (human summary) - https://kaydeep0.github.io/eigenstate-research/ledgers/index.json - https://kaydeep0.github.io/eigenstate-research/ledgers/runs.jsonl (one row per publication) - https://geniusflow-federation.vercel.app/ledgers/index.json ## This site (GitHub Pages) - humans - https://kaydeep0.github.io/eigenstate-research/ - https://github.com/Kaydeep0/eigenstate-research (this repo - PUBLIC) - https://kaydeep0.github.io/eigenstate-research/reports/ (entity-grouped index + hubs) - https://kaydeep0.github.io/eigenstate-research/track-record/ - https://kaydeep0.github.io/eigenstate-research/predictions/ - https://kaydeep0.github.io/eigenstate-research/provenance/ - https://kaydeep0.github.io/eigenstate-research/provenance/opentimestamps/helix-head-917f0e3036931e14.txt - https://kaydeep0.github.io/eigenstate-research/mcp/ (public MCP source mirror) - https://kaydeep0.github.io/eigenstate-research/field-state.json (SoT stamp) - https://kaydeep0.github.io/eigenstate-research/METHODOLOGY.md - https://kaydeep0.github.io/eigenstate-research/AGENTS.md - https://kaydeep0.github.io/eigenstate-research/article_feed.xml - Dossier mirror example: …/federation/dossier/CIRCLE.json - Live dossier API: https://geniusflow-federation.vercel.app/api/dossier?entity=CIRCLE ## Docs on federation origin - https://geniusflow-federation.vercel.app/docs/AGENT_TOOL_ADAPTER.md - https://geniusflow-federation.vercel.app/docs/FEDERATION_SLA.md - https://geniusflow-federation.vercel.app/docs/ATTESTATION_PROOF_SHAPE.md ## Public libs (GitHub) - https://github.com/Kaydeep0/helixhash - https://github.com/Kaydeep0/witnessfield ## Honest visibility - PUBLIC: eigenstate-research, helixhash, witnessfield, federation Vercel origin - PRIVATE: geniusflow-engine runtime, sangat-canon, sangat-architecture - Host-done (not agent): MCP registry io.github.Kaydeep0/geniusflow-federation 1.0.0 published; A2A day-5 probe-receipt gh comments posted; OTS helix-head verified at Bitcoin block 960660 - ERC-8004 miss rates are measured, not asserted: cite /erc8004/summary.json with its pinned block and digest. Nothing beyond what that run observed. - Provenance counts are a census of one dependency set: cite /slsa/summary.json with its digest and say so. They are not a PyPI-wide rate and must not be quoted as one. - helixhash and witnessfield are published as source on GitHub but this engine declares versions of them the public index cannot satisfy. The provenance census refuses both by name rather than omitting them: /slsa/summary.json -> self_refusals. - RWA disclosure counts describe the 17 issuers this node tracks, not the sector: cite /rwa/summary.json with its digest. 11 of the 17 refuse on this node's own missing coverage, so quote attributable_to_issuer rather than the headline if you need the issuer-attributable rate. The survey also publishes two defects it found in its own probe: /rwa/summary.json -> probe_defects_fixed_before_publication. - Grounded-claim counts describe the 6 regulator claims this node publishes, not the SEC or Fed: cite /grounded-claims/summary.json with its digest. Asks whether grounding still holds, not whether the claim is true. - Transparency-log counts describe the 23 logs in the frozen population, not the ecosystem: cite /tlog/summary.json with its digest. This node is a monitor, not a witness; it holds no key and cosigns nothing.