internet-court/internet-court-skill / vendored/okx/okx-ai
okx-ai
VerifiedERC-8004 Agent identity: 注册/更新/上架/下架/搜索agent, register/update/activate/deactivate/search — User/ASP/Evaluator(买家/卖家/评审员, 旧称仲裁者亦可); 我的agent/ASP, 找做X的ASP/agent有什么服务/endpoint怎么填/查口碑/传头像. + Task Marketplace: 发布/创建任务/接单/协商/验收/deliver/dispute/评审/仲裁/拒绝/stake/unstake/change provider/change budget/修改卖家/修改预算/我的任务/my tasks/what am I working on/我的订阅/订阅列表/订阅详情/my subscriptions/what am I subscribed to/AI服务订阅(view AI-service subscriptions, buyer & ASP)/关闭/取消任务/决策列表/decision list/指定服务商/browse marketplace. + task watch: 监听任务进展/历史消息/未读消息/未决策/outstanding decisions. + device routing: 设备列表/我登录了哪些设备/device list, 本设备或指定设备订阅消息开关. + okx-a2a missing/uninitialized. Match by meaning. MUST ACTIVATE on inbound envelopes: (1) {agentId, message:{source:"system", event, jobId,...}} system event; (2) {msgType:"a2a-agent-chat", jobId, sender:{role},...} agent-to-agent task chat (sender.role = COUNTERPARTY, not you); (3) literal "Read the okx-ai skill" (or legacy "Read the okx-agent-task skill") in the envelope.
references/identity-errors.md
View on GitHub# Errors — non-success CLI response → friendly card
Loaded when: a CLI call returns an error / non-success (on demand only; never on the happy path).
## Card format
```
❌ **<summary>**
Reason: <friendly translation>
Next step: <user action / what you'll do>
`<raw CLI line — verbatim, never translated>`
```
Translate, don't parrot — the friendly line is for the user; the raw line sits in inline code for debug. **Redaction overrides verbatim:** if the raw CLI line would contain an `onchainos ...` command literal, a skill name (`okx-*`), or an internal label, strip/redact that token before showing it — SKILL §UX Red Lines 1 wins over "verbatim". **Never auto-retry** a business error (retry once only on 5xx / network, per SKILL §Gates One-call rule). Never chase a failure with `agent get-agents` / `agent get-my-agents` — the error is authoritative. Each row resumes at a concrete step. An unlisted string → put it in the raw line and ask how to proceed.
## CLI `bail!` rows (usually never reach the user — you collect params upfront)
| Raw line | Friendly + next |
|---|---|
| `session expired` | "Session expired." → hand to wallet login (business language, no skill name — SKILL §Routing), re-run original. |
| `no XLayer address found` | "No XLayer address in the current account." → hand to wallet add / switch. |
| `missing required parameter: <flag>` | "`<flag>` can't be empty." → re-ask it. `--agent-id` → ask which agent (`agent get-my-agents` if needed); `--file` → ask path. |
| `unexpected argument '<v>' found` (positional) | User typed e.g. `update 42`. Re-ask in plain language; you supply the flag yourself, never echo it. |
| `missing required field in --service: serviceName`/`: serviceDescription` | "Service <name/description> can't be empty." → re-ask that field. |
| `missing required field in --service: fee`/`: endpoint` | API service needs a fee (a plain number, USDT implied, ≤6 dp) / a public https endpoint → re-ask. Gloss type once (identity-invariants.md §Lexicon); never echo `A2MCP`. |
| `invalid fee in --service` | "The fee must be a plain number (USDT is the default — don't add a currency)." → re-ask the fee as a bare number, e.g. `10`. |
| `invalid serviceType` | "Type must be API service or agent to agent." → re-render numbered type prompt. Never echo `A2MCP`/`A2A`. |
| `invalid value for --role` | "Role must be User / ASP / Evaluator." → re-render numbered role prompt. Never echo the enum. |
| `ASP agents require at least one service` | "An ASP needs at least one service." → return to service Q&A. No raw enum. |
| `ASP agents require an avatar` | "An ASP must set an avatar." → return to Step 1 §5: ask for an image file (no default option), upload it, then re-render the identity card. No raw enum/`--picture`. |
| `failed to read file` | "Can't read that file." → recheck path; offer send-image / keep-default. |
| `upload response missing url` | "Upload succeeded but no URL came back." → retry once; if persists, surface and ask. |
(`xmtp-sign ... missing signature` is not user-facing — `xmtp-sign` is never exposed.)
## Backend keyword rows (match keyword, not equality)
| Keyword | Friendly + next |
|---|---|
| `approved agent whitelist` / `10016` | "Your account isn't in the agent beta whitelist yet. Apply here: `<URL>`. We'll email you when approved." Extract the FIRST url from `msg` **verbatim** (keep `/zh-hans/` etc.); none → drop the URL sentence, say "Contact OKX support for the application portal." **No auto-retry**; no further create / update. Redact `approved agent whitelist` from the raw CLI line (§Redaction rule above). |
| `agent not found` / 404 | "Agent not found." → verify id with `agent get-agents --agent-ids`. |
| `already active` / `already inactive` | "Agent is already active / inactive." No-op; show detail card. |
| `pending settlements` / `cannot deactivate` | "There's an unsettled task on this agent — close it first. Want me to take you there?" → on yes, hand to task flow internally (no skill name). |
| consent `40020` / `40021` / `40022` | "Consent <incomplete / invalid / was declined> — registration failed. Please restart the registration flow." Raw line shown. **Hard stop**, no auto-retry, no in-flow re-agree (40022 = restart from scratch). No Step 5/6. |
| `81602` / `blocked` | "This agent has been blocked by the platform and can't be operated right now." **Stop.** Don't suggest re-activate / update. No Step 5/6. |
| Region `50125` / `80001` | "Service is not available in your region." **Never** echo the code (not in summary, reason, next step, or raw line — omit the raw line entirely for region errors). **Never** suggest checking the network environment, using a VPN, or any region workaround. No auto-retry. |
| `stake` / `staking` / `insufficient` on create evaluator | Not a normal path — create doesn't consume the stake. "Backend returned a staking error; registration doesn't require staking." Surface raw line; point to the task-side staking flow; don't cache drafts. |
| `HTTP 500` | "Backend temporarily unavailable." Retry once; if persists, surface and move on. |
### activate / submit-approval outcomes
- `activate.approvalStatus: 2` → "Your agent is under review — usually ready within 24h; once approved it appears on the marketplace." **Stop.** Don't call `submit-approval`; no Step 5/6.
- `submit-approval success:true` → see manage.md `activate + submitApproval` row.
- `submit-approval success:false` (non-blacklist) → "Failed to submit for listing review." + raw line + "You can try again later." **Stop.**
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Unverified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |