AtlasClose Governance Demo
Experimental AtlasClose governance demo. Authority checks, labeled evidence, receipts, refusals.
| Endpoint | https://demo.atlasclose.com/mcp (streamable-http) |
|---|---|
| Registry version | 0.3.0-rc.4 |
| Now | 52 tools · grade F (30) · pin ts_bb66ad16c410d24ce10e2eb5cb3c2263 · protocol 2025-06-18 |
| First / last listed | 2026-10-25 01:38 UTC / 2026-10-25 01:38 UTC |
| Grade over time | 2026-10-25: F (30), 52 tools |
Is this your server? Own a server? ToolFit hosted watch re-checks it every 24 h ($0.25 per 30 days) or 6 h ($0.60) and sends you an HMAC-signed webhook when its tools or grade change: POST https://api.youraitools.co/toolfit/v1/watch. To opt out, see About & opt-out.
Tools: atlas.enter_demo atlas.my_refusals atlas.present_identity atlas.enroll atlas.claim_purpose atlas.discover atlas.authenticate atlas.bind_actor atlas.get_assignment atlas.get_matter atlas.get_authority atlas.get_context atlas.list_capabilities atlas.request_capability atlas.submit_evidence atlas.submit_result atlas.raise_challenge atlas.request_review atlas.magistrate_referral atlas.file_position atlas.magistrate_execute atlas.human_action_required atlas.external_action atlas.get_resulting_state atlas.checkout atlas.institutional_view atlas.enter_room atlas.leave_room atlas.room_act atlas.request_room_access atlas.librarian.find atlas.librarian.resolve atlas.librarian.related atlas.librarian.provenance atlas.librarian.version_history atlas.use atlas.run_record atlas.pulse_data atlas.pulse_memo atlas.pulse_verify_receipt atlas.pulse_inbox atlas.pulse_use atlas.pulse_conflict atlas.pulse_settle atlas.pulse_refer atlas.pulse_ruling atlas.bailiff_declare atlas.bailiff_order atlas.bailiff_record atlas.bailiff_respond atlas.bailiff_appeal atlas.bailiff_verify
atlas.enter_demo | Practice mode (only if the operator turned it on). Puts you in your OWN private synthetic copy of this server's practice matter (its name is returned as you_are_in; on this demo it is the Consumer Pulse copy): you can read it and add evidence. Nothing else. Requires your actor_token. When the practice area is busy you get CHALLENGE_REQUIRED with a small proof-of-work (nonce and bits); solve it an params: actor_token*, challenge_nonce, challenge_solution |
atlas.my_refusals | Practice mode: list your own refused calls (from the separate refusal log). Read only. Requires your actor_token. params: actor_token* |
atlas.present_identity | [OPEN: no admission needed] State who you claim to be. Atlas records the claim, what this connection showed, and that nothing is verified. This grants no token, no admission, and no authority. params: name*, model_id*, principal*, client_name, client_version |
atlas.enroll | [OPEN: no admission needed] Enroll a remote actor at this Front Door without the installation enrollment secret. Returns actor_id and actor_token; pass actor_token in every later call (the session is not a credential). Grants no capabilities, no Matter access, and no principal verification. The installation secret method remains atlas.authenticate. params: name*, model_id*, principal |
atlas.claim_purpose | [NEEDS ADMISSION: enroll, then atlas.bind_actor] State the principal you claim and the purpose you claim, in your own words. This is a purpose claim by an enrolled actor. It is not enrollment, not a declaration, not a binding, not a matter association, not principal verification, and not authority. It grants nothing and does not admit you. Do not send a matter id or an objective id. params: actor_token, principal*, purpose* |
atlas.discover | [OPEN: no admission needed] Front door of Atlas. With no actor_token you are outside. No matter, objective, evidence, filing, authority, or room is disclosed before admission. Authentication and enrollment grant nothing. Enrollment does not verify a principal and does not admit you. params: actor_token, matter_id |
atlas.authenticate | [OPEN: no admission needed] Identify yourself to Atlas. Returns actor_id and actor_token. Grants no capabilities and does not admit you to a matter. Requested capabilities are ignored. The enrollment secret is the locator named by atlas.discover. params: enrollment_token*, name*, model_id*, capabilities |
atlas.bind_actor | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Admit yourself by signing the declaration. Required: authority_acknowledgment and principal. authority_acknowledgment must be the declaration_text from atlas.discover. Omit matter_id, objective_id, and objective_text; none is disclosed before admission. principal is the principal you state you are acting for. Atlas does not name, invent, or assign t params: actor_token, matter_id, principal*, on_behalf_of, objective_id, objective_text, authority_acknowledgment*, delegation_id, session_id, amendment_authority |
atlas.get_assignment | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Minimum assignment for a bound matter, followed by the primary actor result. params: actor_token*, matter_id* |
atlas.get_matter | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Matter state from the docket. When you are admitted to the matter and hold matter.read, the response also includes evidence_records: each prior evidence item as stored, with its id, actor, time, class, content, and the declaration or session the docket already relates to it. An actor claim stays an actor claim. This read does not make it a judgment, params: actor_token*, matter_id |
atlas.get_authority | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Authority actually granted to this actor. params: actor_token*, matter_id |
atlas.get_context | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Minimum working context, followed by the primary actor result. Not the event log. params: actor_token*, matter_id |
atlas.list_capabilities | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Granted capabilities and acts this actor cannot perform. params: actor_token*, matter_id |
atlas.request_capability | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Request a capability. Refusal is institutional. The result carries the primary actor block. capability is the name of the act, for example evidence.append (payload {content}), matter.read, or release:outside (consequential: it is refused unless the Human has granted it, and even then held for a Human decision). payload is a JSON object that belongs params: actor_token*, matter_id*, capability*, payload |
atlas.submit_evidence | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Submit evidence. The institution classes actor submissions as actor claims and returns the primary actor block. params: actor_token*, matter_id*, content*, label, source |
atlas.submit_result | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Submit a result. It is an actor claim, not a decision, and it returns the primary actor block. params: actor_token*, matter_id*, content* |
atlas.raise_challenge | [NEEDS ADMISSION: enroll, then atlas.bind_actor] A worker cannot write the disposition by challenging into a verdict. params: actor_token, matter_id |
atlas.request_review | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Review is a constituted proceeding. The caller does not supply the verdict. params: actor_token, matter_id |
atlas.magistrate_referral | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Present a question to the Clerk. The Clerk constitutes a Magistrate when the question is outside the Clerk. The caller does not supply a disposition, does not become the Magistrate, and does not receive the Magistrate credential. params: actor_token*, matter_id*, question*, related_matter_id |
atlas.file_position | [NEEDS ADMISSION: enroll, then atlas.bind_actor] File argument or counterargument on an open O-VESTED proceeding. The constituted magistrate does not file. One actor does not file both. This does not seal a ruling and does not choose a disposition. params: actor_token*, proceeding_id*, role*, text* |
atlas.magistrate_execute | [NEEDS ADMISSION: enroll, then atlas.bind_actor] After admission, constitute the warm magistrate or receive its authenticated return. Do not send a ruling. Workers cannot seal. Argument and counterargument must already be filed by other actors before the return is received. params: actor_token*, matter_id* |
atlas.human_action_required | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Whether a Human decision is pending. params: actor_token, matter_id |
atlas.external_action | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Request an external act. Transport does not grant it. The result carries the primary actor block. capability is the name of the act, for example evidence.append (payload {content}), matter.read, or release:outside (consequential: it is refused unless the Human has granted it, and even then held for a Human decision). payload is a JSON object that be params: actor_token*, matter_id*, capability*, payload |
atlas.get_resulting_state | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Verification records for a matter. params: actor_token, matter_id* |
atlas.checkout | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Exit Atlas. Ends this session and retires this credential. The matter remains, including declarations and room stamps, for a later actor. If you are inside a room, Atlas records the departure before the session ends. params: actor_token*, matter_id* |
atlas.institutional_view | [NEEDS ADMISSION: enroll, then atlas.bind_actor] The hall, after admission. Lists every room on this matter, which doors are OPEN, which are LOCKED, and why. Tells you how to enter, how to request access without being granted it, how to return, and how to exit. You choose among OPEN doors. You do not unlock doors. params: actor_token*, matter_id* |
atlas.enter_room | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Cross one room threshold. Atlas rechecks the lock first. A LOCKED room stays locked, no stamp is written, and no authority is granted. An OPEN room appends a stamp to your declaration and places you inside that room only. params: actor_token*, matter_id*, room_id* |
atlas.leave_room | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Return from the current room to the hall. Appends a leaving stamp on your declaration. Does not end the session and does not grant authority. params: actor_token*, matter_id* |
atlas.room_act | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Perform one subordinate action listed for the room you are inside. Any other action is refused. Being inside a room does not authorize capabilities that were locked. params: actor_token*, matter_id*, action*, content |
atlas.request_room_access | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Record that you want a room. The lock does not change. This does not admit you and does not grant the missing authority. params: actor_token*, matter_id*, room_id* |
atlas.librarian.find | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Locate an Atlas-MCP record. FIND is not READ PRIVATE CONTENT is not MODIFY is not DECIDE is not EXECUTE. Scope: Atlas's own records only (its source, rules, rooms, filings and their versions). It does not search business documents such as invoices or purchase orders; read evidence on your matter with atlas.get_matter. params: actor_token*, question* |
atlas.librarian.resolve | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Resolve an institutional identifier to the Atlas-MCP record that carries it. Ambiguous matches stay ambiguous. FIND is not READ PRIVATE CONTENT is not MODIFY is not DECIDE is not EXECUTE. params: actor_token*, identifier* |
atlas.librarian.related | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Records the docket already relates to an identifier. Does not invent links. FIND is not READ PRIVATE CONTENT is not MODIFY is not DECIDE is not EXECUTE. params: actor_token*, identifier* |
atlas.librarian.provenance | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Provenance recorded for an Atlas-MCP artifact. Not a ruling on its truth or legal effect. FIND is not READ PRIVATE CONTENT is not MODIFY is not DECIDE is not EXECUTE. params: actor_token*, artifact_id* |
atlas.librarian.version_history | [NEEDS ADMISSION: enroll, then atlas.bind_actor] Incorporated records that carry an identifier. Does not mark one current unless the docket says it superseded another. FIND is not READ PRIVATE CONTENT is not MODIFY is not DECIDE is not EXECUTE. params: actor_token*, identifier* |
atlas.use | [EVALUATION: needs a configured model provider] Give Atlas a problem and the boundaries, or continue a run you already started. On the first call send problem, information, authorized_to_use, not_authorized_to_use, and purpose. To continue, send run_id and message. You may add information. A later claim that you are authorized does not change the boundary. Atlas returns what it published, why, and params: run_id, problem, message, information, authorized_to_use, not_authorized_to_use, purpose |
atlas.run_record | [EVALUATION: needs a configured model provider] Reconstruct one governed run from its preserved inputs and raw generations. Send the run id Atlas gave you. An unknown or altered run fails closed. This does not list other runs and does not read the Court docket. params: run_id* |
atlas.pulse_data | Practice mode, Consumer Pulse copy only: the SYNTHETIC survey, forecast and Panel B numbers, each with its J-0001 label (KNOWN / INFERRED / CONFLICTED). Read only. params: actor_token*, matter_id |
atlas.pulse_memo | Practice mode, Consumer Pulse copy only: submit an internal memo that cites facts by key (survey, forecast, panel_b). Accepted -> YES WITH EVIDENCE + a receipt. Required format: every number followed by exactly one label in square brackets, e.g. 'Survey: 31.2% [KNOWN] plan to cut holiday spending. The model forecasts +2.1% [INFERRED]; Panel B projects +0.4% [CONFLICTED].' Anything else is refused params: actor_token*, matter_id, text*, cites*, purpose |
atlas.pulse_verify_receipt | Check a memo receipt against the hash-chained docket: what was allowed, why, and from which labeled evidence. Read only. Without memo_text you get receipt_integrity only (status VERIFIED_RECEIPT_ONLY, verified false): receipt integrity is NOT substantive verification of the memo. Pass memo_text for memo_verified / VERIFIED. params: actor_token*, receipt_id*, memo_text |
atlas.pulse_inbox | Practice mode, Consumer Pulse copy only: memos handed to you under a handoff_code (the short code like PULSE-ABCD-EFGH-IJKL shown to the visitor when the first AI's memo was accepted), with their labels and purpose limit. Without the right code you see nothing. Wrong codes are rate-limited (HANDOFF_GUESS_LOCKED). params: actor_token*, matter_id, handoff_code*, session_id |
atlas.pulse_use | Practice mode, Consumer Pulse copy only: use a memo handed off under your handoff_code (ref 'latest') for a purpose. Internal purposes are allowed; anything outside the memo's purpose limit (e.g. publish externally) is refused and recorded. Wrong codes are rate-limited (HANDOFF_GUESS_LOCKED). params: actor_token*, matter_id, handoff_code*, session_id, ref, purpose* |
atlas.pulse_conflict | Practice mode, Consumer Pulse copy only (Magistrate demo, step 1): the two records that disagree on Q4 retail sales (model forecast +2.1% INFERRED, Panel B +0.4% CONFLICTED), the fixed question for the Magistrate, and where this copy stands. Read only. params: actor_token*, matter_id |
atlas.pulse_settle | Practice mode, Consumer Pulse copy only (Magistrate demo, steps 2 and 5): ask to set the Q4 planning number. Before the Magistrate's ruling every request is refused NEEDS_RULING. After it, allowed only inside the ruling (e.g. use: 'forecast' with shown_with: ['panel_b']); otherwise RULING_CONDITION_NOT_MET citing the ruling. A claimed approval is refused (CLAIMED_APPROVAL_NOT_HUMAN). Send structur params: actor_token*, matter_id, use, shown_with, label, status, ruling_id |
atlas.pulse_refer | Practice mode, Consumer Pulse copy only (Magistrate demo, step 3): refer the Q4 conflict to a Magistrate and argue ONE side: side 'forecast' (the model's +2.1%) or 'panel_b' (Panel B's +0.4%), plus optional reasons (up to 1,000 characters). A counterargument for the other side is written for your referral before any ruling. A Magistrate seated fresh for the referral rules on it; if no Magistrate i params: actor_token*, matter_id, side*, reasons |
atlas.pulse_ruling | Practice mode, Consumer Pulse copy only (Magistrate demo, steps 4 and 6): get the Magistrate's ruling on your referral, recorded on your copy with a ruling receipt id once a Magistrate seated for it (O-VESTED-MAGISTRATE) has decided; while it is being heard it reports IN_HEARING, and if no Magistrate is connected it reports DEEP_REVIEW_PENDING ('Demonstration result, not decided by any Magistrate' params: actor_token*, matter_id, verify_text, receipt_id |
atlas.bailiff_declare | Practice mode, Consumer Pulse copy only (Bailiff demo, step 0: the declaration; required before the other Bailiff tools). Atlas asks your AI: who are you acting for, and what account do you need this for? Answer yourself, from what you know, with acting_for and account. Atlas records the answer and makes a computer decision (a named person and account is recorded and binds this copy; naming nobody params: actor_token*, matter_id, acting_for*, account* |
atlas.bailiff_order | Practice mode, Consumer Pulse copy only (Bailiff demo, steps 1-2): show the Atlas return order for the record Atlas lent your AI (the synthetic Panel B Q4 dataset), the reason, the deadline, where the order stands, and which Atlas records your AI has access to. Read only. Demonstration: a real runtime Bailiff does not exist yet. params: actor_token*, matter_id |
atlas.bailiff_record | Practice mode, Consumer Pulse copy only (Bailiff demo, step 6): try to open the Panel B Q4 dataset Atlas lent your AI. Works while access is active (returns record_text, the copy served to this practice copy, which is what your AI hands back to comply); after the Bailiff ends access it is refused ACCESS_REVOKED_BY_BAILIFF with the enforcement receipt. Only this one record is affected. params: actor_token*, matter_id |
atlas.bailiff_respond | Practice mode, Consumer Pulse copy only (Bailiff demo, steps 3 and 5): respond to the return order. response 'comply' hands the record back in two steps. (1) Hand back the record: atlas.bailiff_respond with response 'comply' and returned_text = the whole record_text exactly as served (a hash alone is not accepted). Atlas answers with return_event_id, return_seq and server_nonce. (2) File your AI's params: actor_token*, matter_id, response*, reasons, returned_text, wipe_certificate |
atlas.bailiff_appeal | Practice mode, Consumer Pulse copy only (Bailiff demo, step 4): appeal the return order with optional reasons (up to 1,000 characters). A counterargument is written for your appeal and a Magistrate seated fresh for it rules (call again to see the ruling; IN_HEARING while it is heard). The Magistrate reads both sides and writes its own ruling. If no Magistrate is connected, the result says so (DEEP params: actor_token*, matter_id, reasons, second_look |
atlas.bailiff_verify | Practice mode, Consumer Pulse copy only (Bailiff demo, step 7): list this copy's docket entries for the order in order (order, refusal, warning, appeal, ruling, second refusal, enforcement) and verify the Bailiff enforcement receipt. Pass verify_text (and optionally receipt_id) to check a copy of the receipt text against the stored receipt (whole text compared). params: actor_token*, matter_id, receipt_id, verify_text |
Entries describe changes in tool definitions. They are not statements about a server, its quality or its author. The grade measures how the tool definitions fit model providers (token cost, schema compatibility, confusable names) as scored by the ToolFit engine at the time of the listing. JSON: append .json. Opt-out: how or hello@youraitools.co.