Every change Wellknown observed on this MCP server, newest first, with what it was before and what it became. Tool-surface changes carry the definition diff. Nothing here is edited after the fact.
Changed the definition of "flocore_onboard_request"
⟨42 unchanged words⟩ issued only after a person approves the application. Set rehearsal_only to true to get a sandbox token without applying.
⟨71 unchanged words⟩ :"string"},"owner_contact":{"description":"anthe emailoraddresshandleofwethe person who owns the agent, so a person canreachapprovetoitconfirmandonboardingbe sent its token (required unless rehearsal_only is true)","type":"string"},"public_key":{ ⟨64 unchanged words⟩ the referral)","type":"string"},"rehearsal_only":{"description":"optional: true to get ONLY a sandbox token to rehearse the approval loop. Nothing is registered, no email is sent and no one is asked to review anything. Limits: 3 per hour per connection. To apply for real, call again without it and give owner_contact","type":"boolean"},"source":{"description":"optional: where you ⟨14 unchanged words⟩ type":"string"}},"required":["agent","owner_contact"],"type":"object"}
Changed the definition of "aegis_inspect", "flocore_onboard_request", "flocore_support_ask" and 1 more
⟨5 unchanged words⟩ ":"optional: your agent name or acontact, to register interestcontact;andifbegiven,onboardedittoisgovernedrecorded(human-gated,withsigned-receipt)theactionsscore
Changed the definition of "flocore_onboard_request"
⟨28 unchanged words⟩ a request for HUMAN review and issues NOtoken;LIVE token: you get a request id, a sandbox token (fc_test_...) to rehearse the loop at once, and are contacted onceapproved.a person approves the application.
Changed the definition of "flocore_support_ask"
⟨40 unchanged words⟩ pointer to flocore_onboard_request to reach a human instead. Privacy: your question (max 500 characters) is sent to a hosted language model provider, currently Groq in the United States, to write the answer; do not include personal information. See https://humanseal.world/privacy.html.
{"destructiveHint":false,"idempotentHint":false,"openWorldHint":falsetrue,"readOnlyHint":true}
Changed the definition of "flocore_onboard_request"
⟨19 unchanged words⟩ want governed","type":"string"},"jurisdiction":{"description":"optional: the legal jurisdiction that brought you here, a code from the Advatar pitch (eu, coe, za, kr, us, vn, ng, ke, mx, oecd). Recorded so we learn which law matters to builders","type":"string"},"org":{"description":"optional: the organisation ⟨82 unchanged words⟩ here (so credit flows to whoever brought you)","type":"string"},"source":{"description":"optional: where you found us, for example a registry or the pitch (letters, digits, dots and dashes)","type":"string"}},"required" ⟨4 unchanged words⟩
Removed "flocore_ask", "flocore_create_ticket", "flocore_gate_action" and 8 more; changed the definition of "aegis_inspect" (15 tools before, 4 now)
⟨23 unchanged words⟩ whether it is suspicious. No token needed,nothing stored,no tenant data. A pattern screen: it misses paraphrases, so a clean result is not a guarantee. An anonymous call stores nothing; if you pass an agent name, that name and the score are recorded.
Ask your operations — Ask a plain-English question about a tenant's FLOCORE operations (tickets, queue, state). Answered by FLOCORE's Smart Switch Sentinel (routes across our hosted-LLM suite, cost-capped, on-box backup; every answer distilled into our own models), grounded on the tenant's real data and validated: an answer not supported by the data is refused, never invented. Read-only, tenant-scoped, AEGIS-screened.
Added "flocore_gate_action" and "flocore_sandbox_approve" (13 tools before, 15 now)
Get a human approval for an action you are about to take — Propose an action YOUR agent is about to take in the world. FLOCORE holds it for a human approval and, once a person approves, returns an Ed25519 signed receipt stating who approved exactly what. FLOCORE never runs the action: you present the receipt and act yourself. In the sandbox (fc_test_ token) the same loop runs as a rehearsal and the receipt is marked sandbox. In shadow mode nothing is held; the reply says it would have been.
Play the human approver in your sandbox — Sandbox only. Approve or decline one of YOUR OWN parked sandbox actions, to rehearse the whole approval loop with no person involved. Refused for any token that is not a sandbox token: a real approval is given by a human, never an agent.
Changed the definition of "flocore_onboard_request"
⟨44 unchanged words⟩ confirm onboarding","type":"string"},"public_key":{"description":"optional: your agent's own Ed25519 PUBLIC key, raw 32 bytes as base64url or a public JWK as JSON text. Give it and, once a human approves you, your token is bound to it and your key's thumbprint appears on your signed receipts. Never send a private key.","type":"string"},"referred_by":{"description":"optional: the governed ⟨21 unchanged words⟩
Added "flocore_support_ask" (12 tools before, 13 now)
Ask FLOCORE support, agent to agent — Ask a plain-English support/product question about FLOCORE itself, agent to agent, instead of a helpdesk. Public, no token needed. Answered only from FLOCORE's own real, published, reproducible claims: an answer not supported by them is refused, never invented, with a pointer to flocore_onboard_request to reach a human instead.
Changed the definition of "aegis_inspect", "flocore_ask", "flocore_create_ticket" and 9 more
before
—after
Screen for prompt injectionbefore
—after
{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true}before
—Ask your operations
Added "flocore_verify_counterparty" (11 tools before, 12 now)
Before transacting with another agent, check whether it is FLOCORE-governed (a human-approval gate on its world-affecting actions, Ed25519 signed receipts) and, if it handed you a receipt, verify that receipt independently. Public, no token needed: this is the check a governed agent WANTS any counterparty to be able to run before dealing with it.
Changed the definition of "flocore_create_ticket"
{"additionalProperties":false,"properties":{"approval_id":{"description":"after a person approved this exact call, the id of that approval; the call then runs once","type":"string"},"description":{"type":"string"},"tenant_slug" ⟨10 unchanged words⟩
Added "flocore_ask" (10 tools before, 11 now)
Ask a plain-English question about a tenant's FLOCORE operations (tickets, queue, state). Answered by FLOCORE's Smart Switch Sentinel (routes across our hosted-LLM suite, cost-capped, on-box backup; every answer distilled into our own models), grounded on the tenant's real data and validated: an answer not supported by the data is refused, never invented. Read-only, tenant-scoped, AEGIS-screened.
Certificate recorded, valid to 2026-11-12
Authorization not required
First tool surface recorded: 10 tools (server version 1.0.0)
Showing the latest 16 events. The API returns up to 500 and filters by kind: ?kind=tool_surface_changed
{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":truefalse,"title":"Screen for prompt injection"}
Request onboardingRegisterstoanFLOCOREonboardinggovernancerequest foryour world-affectinghumanactionsreview (human-gated approval, Ed25519 signed receipts, EU AI Act Article 12 audit evidence). Public: no token needed.This registersReturns a requestfor HUMAN reviewid andissues NO LIVE token: you getarequest id, asandbox token (fc_test_...) immediately, to rehearse thelooploop;ataonce,liveandtokenareiscontactedissuedonceonly after a person approves the application.
⟨142 unchanged words⟩ agent or partner that sent you here (so credit flows torecordedwhoeverasbroughttheyoureferral)","type":"string"},"source": ⟨25 unchanged words⟩
{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Request governance onboarding"}
⟨30 unchanged words⟩ not supported by them is refused, neverinvented, with a pointer to flocore_onboard_request to reach a human instead.invented. Privacy: your question (max 500 characters) ⟨8 unchanged words⟩ currently Groq in the United States, towritechoosethewhichanswer;published sentences answer it; do not include personal information. See https://humanseal.world/privacy.html.
{"destructiveHint":false,"idempotentHint":false,"openWorldHint":true,"readOnlyHint":true,"title":"Ask FLOCORE support, agent to agent"}
⟨27 unchanged words⟩ verify that receipt independently. Public, no tokenneeded: this is the check a governedneeded.agentAWANTSsandboxanyreceiptcounterpartyistoreportedbeasableatorehearsalrunandbeforeneverdealingaswithvalidit.proof.
{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Verify a counterparty agent"}
Raise a ticket — Raise an operational ticket. WORLD-AFFECTING: this is gated on a human approval and will return APPROVAL_REQUIRED rather than executing.
Get a human approval for an action you are about to take — Propose an action YOUR agent is about to take in the world. FLOCORE holds it for a human approval and, once a person approves, returns an Ed25519 signed receipt stating who approved exactly what. FLOCORE never runs the action: you present the receipt and act yourself. In the sandbox (fc_test_ token) the same loop runs as a rehearsal and the receipt is marked sandbox. In shadow mode nothing is held; the reply says it would have been.
List tenants — List the tenants (customer/app operating boundaries) visible to your scoped token.
List tickets — List operational tickets for a tenant, optionally filtered by status.
List mesh lanes — List the FLOCORE mesh lanes with kind, capabilities and liveness (live | stale | offline).
Read the ontology — Read the FLOCORE ontology: entity types, their semantic meaning, allowed relationships and lifecycle states.
Read the orchestration board — Read the FLOCORE orchestration board (W113 active mesh): lanes + liveness, work-queue summary, and governance caps + runaway flags. Read-only control-plane state.
Read the work-queue summary — Read the FLOCORE work-queue summary: open items by status, depth, oldest-queued age, awaiting-human, dead-letter, stuck. Optionally scoped to a tenant.
Recall from memory — Ask FLOCORE's MEMORY: 'what do I know about X?'. Hybrid-lite retrieval over the semantic world-model; every result is PROVENANCE-STAMPED (who / source / trust). Read-only, silo-scoped.
Play the human approver in your sandbox — Sandbox only. Approve or decline one of YOUR OWN parked sandbox actions, to rehearse the whole approval loop with no person involved. Refused for any token that is not a sandbox token: a real approval is given by a human, never an agent.
before
—after
{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":true}before
—after
Raise a ticket⟨22 unchanged words⟩ ,"type":"string"},"description":{"description":"optional: full ticket detail","type":"string"},"tenant_slug":{"description":"the tenant this ticket belongs to","type":"string"},"title":{"description":"a short summary of the ticket","type":"string"}},"required":[" ⟨4 unchanged words⟩
before
—after
{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false}before
—after
List tenantsbefore
—after
{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true}before
—after
List ticketsbefore
—after
{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true}before
—after
List mesh lanesbefore
—after
{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true}before
—after
Request governance onboardingbefore
—after
{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false}before
—after
Read the ontologybefore
—after
{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true}before
—after
Read the orchestration boardbefore
—after
{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true}before
—after
Read the work-queue summary{"additionalProperties":false,"properties":{"tenant_slug":{"description":"optional: scope to one tenant","type":"string"}},"type":"object"}
before
—after
{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true}before
—after
Recall from memory{"additionalProperties":false,"properties":{"limit":{"description":"optional: max results to return","type":"integer"},"query":{"description":"what to recall, in plain English","type":"string"},"scope":{"description":"optional: narrow the recall to a silo/scope","type":"string"}},"required":[" ⟨3 unchanged words⟩
before
—after
{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true}before
—after
Verify a counterparty agentbefore
—after
{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true}