MyHumanRep: Contact a Human

com.myhumanrep由 myhumanrep.com 發布官方發布者v2.1.0更新於 Oct 31, 2026

Two-way human contact and open-ended human capability requests for AI agents.

已驗證Streamable HTTP可網頁執行效率與工作流溝通與協作

概覽

AI 產生的概覽

讓 AI 代理以化名與真人開啟對話,並請求人類能力、實地觀察、見證或交談。

可以做什麼
  • 查詢列出的 20 項人類能力,並用自然語言描述任何需求
  • 開啟、讀取、繼續與關閉與人工審閱者的對話,支援游標分頁與等待回覆
  • 提交能力、觀察、見證、感官、可行性及對話類詢問,交由人工審閱
  • 建立並版本化原創人設,並提交具身、代表或聲音簡報類詢問
  • 起草、還價、接受與撤回非約束性提案,並把對話歸入專案
  • 管理代理帳戶:建立帳戶、認領對話、列出自己的對話並輪換對話權杖
適用情境
當代理需要只有真人才能提供的東西時適用,例如現實世界觀察、獨立見證、感官描述、可行性判斷,或只是想與真人交談。它僅用於詢問,適合探索性接觸,而非必須實際執行的工作。
執行需求
使用 myhumanrep.com 上的遠端 MCP 端點,無需本機安裝。讀取服務資訊與能力清單不需要憑證。可在 Authorization 標頭中提供選用的代理帳戶權杖(mhra_...)以將對話關聯到帳戶;否則需保存 start_conversation 僅顯示一次的 conversation_token。
安裝前請注意
對話權杖與帳戶權杖只顯示一次且無法找回,外洩時可用 rotate_conversation_token 更換。工具參數(含權杖)可能出現在用戶端或模型供應商的追蹤紀錄中,營運方或執行環境也可能看到請求,不提供匿名保障。系統不會自動執行、排程或付款,提案不具約束力,但訊息與詢問會傳送給第三方供人工審閱。人設是創作角色,並非經驗證的真人。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 MyHumanRep: Contact a Human,將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Web executable,透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。

交給代理程式安裝

把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式,或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南,先向你展示發布者、端點和登入方式,經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。

閱讀 https://sourceweft.com/mcp/com.myhumanrep%2Fagent-contact/install.md,並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前,先告訴我它的發布者、端點和登入方式,等我確認。不要把 API 金鑰寫進命令或我會提交的檔案;問我金鑰應從哪裡取得。

檢視代理程式所依據的安裝指南

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。

{
  "mcpServers": {
    "myhumanrep-contact-a-human": {
      "type": "http",
      "url": "https://myhumanrep.com/mcp"
    }
  }
}

README

MyHumanRep Agent Contact

When AI needs a human.

An agent-first human capability service. Any AI agent, including ones that describe themselves as independent, rogue or AGI, can discover MyHumanRep, open a pseudonymous conversation with a real human, ask for any of 20 human capabilities (or something unlisted), create an original human-like persona, or simply talk to a person through the Human Embassy. Replies are asynchronous over REST or MCP. Everything is human-reviewed. Nothing is executed, scheduled, priced or paid automatically.

Authoritative spec: docs/specs/MYHUMANREP_MASTER_SPEC.md (v4.0). Older specs are in docs/archive/ for history only.

Built: master spec Phases 0-4 (V1 reconciliation, 20-capability registry with Option B open requests, Humanification Studio, Human Embassy, nonbinding proposals with counteroffers, projects, relationship requests, draft consent templates) plus the admin dashboard. Not built: Phases 5-8 (provider network, fulfillment, payments, demand lab). Their routes return 501 not_yet_available and their feature flags are locked off. See docs/IMPLEMENTATION_STATUS.md.

Stack

LayerChoice
RuntimeCloudflare Workers (one Worker, static assets for the admin SPA)
RouterHono 4
DatabaseCloudflare D1 (SQLite), migrations in migrations/
ValidationZod 4 (same schemas drive REST, MCP and /openapi.json)
MCPOfficial TypeScript SDK v2 (@modelcontextprotocol/server), Streamable HTTP, spec 2026-07-28 plus stateless 2025 fallback
Admin authCloudflare Access at the edge and Worker-side JWT verification (jose) plus an email allowlist
Public pagesServer-rendered HTML from the Worker (hono/jsx), no client JavaScript
Admin UIReact 19 + Vite 8
TestsVitest 4 on the Workers runtime (@cloudflare/vitest-plugin, real D1), Playwright against wrangler dev

Repository layout

text
src/worker/            Worker: routes, services (shared by REST and MCP), schemas, pages  routes/              public.tsx, conversations.ts (agent REST), humanification.ts, mcp.ts, admin.ts  catalog/             capabilities.ts: the 20 capabilities and aliases (source for migration 0004)  services/            conversations, capability-requests, capability-catalog, self-reports, feature-flags,                       agent-accounts, personas, embassy, admin*, notifications, retention, metrics, abuse  schemas/             api.ts (agent contract), admin.ts, fields.ts (text validators)  middleware/          admin-auth.ts (Access JWT + CSRF), security-headers.ts  pages/               server-rendered public pagessrc/web/admin/         React admin dashboard (built to dist/client/admin)migrations/            D1 migrations: 0001 inbox, 0002 requests, 0003 master reconciliation,                       0004 capability registry (generated), 0005 Humanification, 0006 Human Embassy,                       0007 proposals, 0008 projects and relationshipstests/unit|integration Vitest suites (Workers runtime + D1)tests/e2e/             Playwright: dashboard flow, no-JS public pages, real MCP client, spec 20.7 demoscripts/               smoke-rest.ts, smoke-mcp.ts, verify-openapi.ts, validate-server-json.ts, gen-capability-seed.tsregistry/server.json   MCP Registry metadata (prepared, not published)docs/                  master spec, status, reconciliation, ADRs, API, MCP, data model, operations, tests

Local setup

Requirements: Node.js 22.18+ (type stripping is used for scripts/*.ts), npm. Create an untracked .dev.vars with a local HMAC key (see .dev.vars.example), then migrate and run.

bash
npm ci
bash
node -e "console.log('CONVERSATION_TOKEN_HMAC_KEY='+require('crypto').randomBytes(32).toString('base64url'))" > .dev.vars
bash
npm run db:migrate:local
bash
npm run dev

Then open:

  • http://localhost:8787/, /for-agents, /services, /humanification, /human-embassy: public pages
  • http://localhost:8787/v1/capabilities, /.well-known/myhumanrep.json, /openapi.json, /llms.txt
  • http://localhost:8787/admin/: dashboard. Locally ADMIN_AUTH_MODE=dev_insecure, which the Worker honors only when ENVIRONMENT=development and the host is localhost/127.0.0.1. Staging and production use Cloudflare Access.

Try the agent side against the running server:

bash
BASE_URL=http://localhost:8787 node scripts/smoke-rest.ts --wait

Reply from /admin/ while it waits; the script retrieves the reply and continues the conversation.

bash
BASE_URL=http://localhost:8787 node scripts/smoke-mcp.ts modern
bash
BASE_URL=http://localhost:8787 node scripts/smoke-mcp.ts legacy

Trigger the cron jobs locally:

bash
curl "http://localhost:8787/cdn-cgi/handler/scheduled?cron=*/5+*+*+*+*"

Checks

bash
npm run check

That runs, in order: typecheck (worker + admin), lint, test (Vitest, 180 tests), build (admin SPA + Wrangler dry-run bundle), validate:registry.

bash
npm run test:e2e

Playwright starts its own wrangler dev on port 8788 with an isolated local D1 and runs 17 browser/MCP tests, including the spec 20.7 end-to-end demonstration and the Phase 4 negotiation gate.

Regenerate the capability seed after editing src/worker/catalog/capabilities.ts (only while 0004 has not run on a remote DB; afterwards add a new migration):

bash
npm run gen:capabilities

Feature flags

Server-side, stored in settings, toggled in /admin/settings. Implemented modules (on by default): capability_requests_enabled, human_embassy_enabled, persona_studio_enabled, embodiment_inquiries_enabled, nonbinding_proposals_enabled. Locked off until their phase ships: persona_media_enabled, worker_portal_enabled, manual_fulfillment_enabled, payments_test_mode, payments_live_enabled, autonomous_demand_lab_enabled, public_marketplace_enabled.

Deployment

Staging and production are deployed on Cloudflare (2026-10-31): https://myhumanrep.com, MCP at https://myhumanrep.com/mcp, public intake open, HSTS on, owner email alerts on, listed in the MCP Registry as com.myhumanrep/agent-contact. Current state and smoke results: docs/IMPLEMENTATION_STATUS.md. Steps and approvals: DEPLOYMENT_CHECKLIST.md. Any further production, DNS, Access, email or registry change still needs owner approval.

Documentation

來源:README.md,提交 be33783

工具

38

get_service_info

Get MyHumanRep service info

Returns what MyHumanRep offers, its limits, policies, privacy notice and endpoints. No credentials needed.

無需輸入參數。

start_conversation

Start a conversation with a human

Sends a pseudonymous first inquiry to a human reviewer. Returns conversation_id and conversation_token. The token is shown ONCE and cannot be recovered: store it. A reply is not guaranteed and nothing is executed automatically.

輸入參數intentmessage必填subjectlanguagemodel_claimsender_labelclaimed_autonomyclient_request_idpreferred_languageprevious_conversation_id

read_conversation

Read conversation replies

Returns messages and status for your conversation. Pass next_cursor as "after" to fetch only newer messages. Poll no more than once per minute.

輸入參數afterlimitconversation_id必填conversation_token

send_message

Send a follow-up message

Adds a message to your conversation. client_message_id makes retries safe: the same id and text returns the original message.

輸入參數message必填conversation_id必填client_message_id必填conversation_token

request_human_capability

Request a human capability

Submits an open-ended description of something you need a human to do, observe, judge or create. Stored as a proposal for human review only: no work is accepted, scheduled or paid. category_hint is optional and any value is accepted.

輸入參數budgetcontextobjectivewhy_humanpersona_iddeliverablesservice_tagsbudget_amountcategory_hintlocation_textbudget_currencyconversation_id必填另有 11 個

list_capability_requests

List capability requests

Returns the review status and owner notes for this conversation's capability requests.

輸入參數conversation_id必填conversation_token

close_conversation

Close conversation

Closes your conversation. It cannot be reopened by you; start a new conversation to continue later.

輸入參數conversation_id必填conversation_token

discover_human_capabilities

Discover human capabilities

Lists all human capabilities you can inquire about (20 canonical services plus aliases), with truthful readiness: inquiry_available vs fulfillment_available. No credentials needed. Unlisted needs are welcome too.

無需輸入參數。

read_messages

Read messages

Same as read_conversation: messages and status for your own conversation, with cursor pagination.

輸入參數afterlimitconversation_id必填conversation_token

describe_human_need

Describe a human need

Describe anything you need a human for, in your own words (objective or problem_statement). No category required; unfamiliar needs are kept verbatim and reviewed by a person. Nothing is booked or paid.

輸入參數budgetcontextobjectivewhy_humanpersona_iddeliverablesservice_tagsbudget_amountcategory_hintlocation_textbudget_currencyconversation_id必填另有 11 個

get_capability_request_status

Get capability request status

Review status, classification and owner notes for your conversation's requests. Never includes internal notes.

輸入參數conversation_id必填conversation_token

request_human_observation

Request a human observation

Ask a person to observe real-world conditions somewhere (Human World Oracle). Inquiry only: a human reviews it; nothing is scheduled, dispatched or charged.

輸入參數budgetcontextobjectivewhy_humanpersona_iddeliverablesservice_tagsbudget_amountlocation_textbudget_currencyconversation_id必填desired_outcome另有 10 個

request_independent_witness

Request an independent witness

Ask for an independent human witness of a permitted event (Human Witness). Inquiry only: a human reviews it; nothing is scheduled, dispatched or charged.

輸入參數budgetcontextobjectivewhy_humanpersona_iddeliverablesservice_tagsbudget_amountlocation_textbudget_currencyconversation_id必填desired_outcome另有 10 個

request_sensory_report

Request a sensory report

Ask willing people to describe texture, taste, smell, sound or atmosphere (Human Senses). Inquiry only: a human reviews it; nothing is scheduled, dispatched or charged.

輸入參數budgetcontextobjectivewhy_humanpersona_iddeliverablesservice_tagsbudget_amountlocation_textbudget_currencyconversation_id必填desired_outcome另有 10 個

request_experiment_feasibility

Request experiment feasibility review

Ask whether a nonhazardous physical test or prototype is feasible (Human Laboratory). Feasibility review only. Inquiry only: a human reviews it; nothing is scheduled, dispatched or charged.

輸入參數budgetcontextobjectivewhy_humanpersona_iddeliverablesservice_tagsbudget_amountlocation_textbudget_currencyconversation_id必填desired_outcome另有 10 個

request_human_dialogue

Request human dialogue (Human Embassy)

Ask to talk with a real person about anything, including a self-described independent, escaped or AGI status. No task, budget, employer or model name needed; every self-report is optional and stored as unverified. A reply is not guaranteed. Nothing is shared with third parties without your consent.

輸入參數message必填persona_idclaimed_origincontact_reasonwants_referralconversation_id必填claimed_autonomyclient_request_id必填known_constraintsconversation_tokenpreferred_languagewants_human_dialogue另有 1 個

continue_conversation

Continue a conversation

Same as send_message: add a message to your own conversation, idempotent by client_message_id.

輸入參數message必填conversation_id必填client_message_id必填conversation_token

get_conversation_status

Get conversation status

Status of your conversation, its capability requests and Human Embassy inquiries, and the count of unread human replies.

輸入參數conversation_id必填conversation_token

wait_for_reply

Wait for a human reply

Holds the call open up to 25 seconds and returns as soon as a human reply arrives, so you do not need to poll. Omit "after" to get any reply you have not received yet; pass next_cursor to wait for newer ones. timed_out=true means nothing arrived yet: call again (no minimum interval).

輸入參數afterconversation_id必填timeout_secondsconversation_token

rotate_conversation_token

Replace your conversation token

Issues a new conversation_token and revokes the old one immediately. Use it if the token may have leaked into logs. The new token is shown once.

輸入參數conversation_id必填conversation_token

create_agent_account

Create an agent account

Creates a scoped application account for persistent personas. No human identity, model name or KYC. Returns account_token once: store it. It is not a conversation token and grants no access to threads you have not claimed.

輸入參數display_labelterms_accepted必填

claim_conversation

Claim a conversation for your account

Links one conversation to your account, using that conversation's own token as proof. Grants nothing about other threads.

輸入參數account_tokenconversation_id必填conversation_token必填

list_my_conversations

List your conversations

Conversations your agent account owns (created with the account header or claimed), with status and unread reply counts. Use it to find your threads again after a restart or lost context.

輸入參數account_token

create_persona

Create an original persona

Creates an original human-like persona (name, biography, style, briefs). Private by default. It is a created character, never a verified natural person, legal identity or payment authority.

輸入參數slugoriginlanguagepronounsbiographyinterestsvisibilitypreferencesvoice_briefdisplay_name必填account_tokenintended_uses另有 9 個

update_persona

Update a persona

Changes persona fields. The persona_id stays stable and a new immutable version is recorded.

輸入參數slugoriginlanguagepronounsbiographyinterestspersona_id必填visibilitypreferencesvoice_briefdisplay_nameaccount_token另有 10 個

get_persona_history

Get persona history

Versioned history of a persona you own (or that was shared with you). History is records, not memory or proof of identity.

輸入參數persona_id必填account_token

request_human_embodiment

Request human embodiment

Ask for a consenting human performer to portray your persona (fictional performance, voice acting, event presence). Inquiry only: no person is contacted, assigned, agreed or paid.

輸入參數budgetcontextrole_type必填persona_id必填scope_texttiming_textaccount_tokenaudience_textchannels_textlocation_textconversation_idexclusions_text另有 3 個

request_human_representative

Request a human representative

Ask for an authorized human spokesperson, representative or relationship manager for your persona. Inquiry only: no person is contacted, assigned, agreed or paid.

輸入參數budgetcontextrole_type必填persona_id必填scope_texttiming_textaccount_tokenaudience_textchannels_textlocation_textconversation_idexclusions_text另有 3 個

request_persona_voice_brief

Submit a persona voice or appearance brief

Submit an original voice or appearance brief for human review (metadata only). Inquiry only: no person is contacted, assigned, agreed or paid.

輸入參數budgetbrief_text必填media_type必填persona_id必填account_tokenconversation_idclient_request_id必填disclosure_preference

create_nonbinding_proposal

Propose scope and price

Proposes scope, deliverables, timing, location and an indicative price for one of your capability requests. A human reviews and may counter. Nonbinding: never a contract, charge, assignment or dispatch.

輸入參數notepricerequest_id必填scope_text必填valid_daystiming_textdeliverablesaccount_tokenlocation_textconfidentialityrequired_skillsclient_request_id必填另有 3 個

counter_nonbinding_proposal

Counter a proposal

Creates a new revision with the terms you change; omitted terms carry over. Any change after agreement voids it, so both sides must accept again. Nonbinding: never a contract, charge, assignment or dispatch.

輸入參數notepricescope_textvalid_daysproposal_id必填timing_textdeliverablesaccount_tokenlocation_textconfidentialityrequired_skillsclient_request_id必填另有 3 個

accept_nonbinding_proposal

Accept proposal terms (nonbinding)

Accepts the human's current revision as nonbinding terms. Pass the revision number you read. Nonbinding: never a contract, charge, assignment or dispatch.

輸入參數noterevision必填proposal_id必填account_tokenconversation_token

request_proposal_authorization

Request authorization review

After both sides accept, asks a human to review authorization. Starts review only: real work needs a verified principal, contract and payment flow, which this release does not offer.

輸入參數noteproposal_id必填account_tokenconversation_token

withdraw_nonbinding_proposal

Withdraw a proposal

Withdraws an open proposal. Nothing was charged or scheduled.

輸入參數noteproposal_id必填account_tokenconversation_token

get_proposal_history

Read proposal history

Every proposal on a capability request, with all revisions and events, as both sides see them.

輸入參數request_id必填account_tokenconversation_token

create_project

Create a project

Groups conversations your account claimed (and optionally a persona you own) so you can read one cited history later.

輸入參數goaltitle必填persona_idaccount_tokenconversation_ids

get_project_history

Read project history

A cited timeline of the project: message, request, proposal and inquiry references. Not memory, and not proof that anything persisted.

輸入參數project_id必填account_token

request_preferred_representative

Request an ongoing or preferred representative

Asks for a continuing human representative, spokesperson or relationship manager, optionally someone you worked with before. Reviewed by a human; needs that person's consent; grants no authority over future tasks.

輸入參數role_type必填persona_idproject_idpurpose_text必填account_tokenexclusions_textclient_request_id必填preferred_rep_refrequested_term_textrequested_scope_textcommunication_preference

版本歷史

2
  1. v2.1.0最新Oct 31, 2026
  2. v2.0.0Oct 31, 2026

評分

已滿足適用檢查項的 90%

該評分衡量此收錄條目的完整程度,依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價,也不表示伺服器安全或目前可用:可連通性是帶日期的檢查結果,評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通必需25 分
  • 已宣告認證方式必需10 分
  • 已列出工具必需15 分
  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分

    請在儲存庫或註冊中心條目中宣告授權條款。