← 返回 Skills 市场
mrtdlgc

revert.wtf Agent API Skill

作者 mrtdlgc · GitHub ↗ · v0.1.0 · MIT-0
cross-platform ✓ 安全检测通过
18
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install revertwtf-agent-api
功能描述
Use revert.wtf safely from agents, tools, browsers, or integrations through bounded HTTP APIs, MCP tools, and lightweight package subpaths without downloadin...
使用说明 (SKILL.md)

revert.wtf Agent API Skill

Use this skill when an agent needs to explain an EVM/RPC/provider/wallet/AA/x402 error, resolve a selector, search the catalog, or integrate revert.wtf into a product.

Default Choice

  • Prefer MCP when the agent runtime supports MCP.
  • Prefer the website HTTP API when calling the hosted service from an agent or app backend.
  • Prefer package subpaths for local code.
  • Never import the full catalog-backed parser into browser/client bundles for one-off searches.

MCP Tools

Use bounded tools:

  • explain_error for raw errors, traces, JSON strings, revert bytes, and x402 payloads.
  • search_catalog for catalog discovery. It returns summaries with limit, offset, totalMatches, and hasMore.
  • get_error only after search_catalog returns a specific id.
  • lookup_selector for 4-byte selectors.
  • decode_revert_data for raw revert bytes.
  • explain_aa_error for AAxx and EntryPoint failures.
  • search_blockscout_chains and get_blockscout_chain for Blockscout registry checks.

Do not request or construct an all-catalog dump. Page through summaries only when the user explicitly needs breadth.

Hosted HTTP API

POST JSON to the public web API from a server/agent context:

curl -s https://revert.wtf/api/explain \
  -H "content-type: application/json" \
  -d '{"raw":"{\"code\":-32000,\"message\":\"nonce too low\"}"}'

Useful endpoints:

  • POST /api/explain with { "raw": "..." }
  • POST /api/revert-decode with { "data": "0x...", "abiText": "[...]" }
  • POST /api/aa-decode with { "raw": "AA23 reverted or OOG" }
  • POST /api/selector with { "selector": "0x08c379a0" }
  • POST /api/search with { "query": "AA23", "limit": 20, "offset": 0 }

Inputs are capped for parsing. Do not paste private keys, seed phrases, access tokens, unreleased exploit details, or personal data.

Use /api/search for discovery and /api/explain for a specific raw failure. Search returns summaries; fetch or explain the selected item instead of asking for broad payloads.

Package Imports

Use precise subpaths:

import { explain } from "@revertwtf/parser/explain"; // server/API path
import { decodeRevertData } from "@revertwtf/parser/decode";
import { describePanic } from "@revertwtf/catalog/panic";
import { listKnownAACodes } from "@revertwtf/aa/parse";
import { searchCatalog } from "@revertwtf/search"; // server-side SQLite FTS

Browser/client bundles should use small helpers or call an API route. Full-data surfaces such as @revertwtf/catalog, @revertwtf/parser/explain, and @revertwtf/aa/explain belong on a server/API boundary.

For browser apps, prefer the lightweight SDK:

import { createRevertClient } from "@revertwtf/client";

const revert = createRevertClient();
const result = await revert.searchCatalog({ query: "x402 payment required", limit: 5 });
安全使用建议
Safe to install for revert.wtf-assisted blockchain error debugging. Avoid sending secrets or sensitive transaction details to hosted APIs, and prefer the bounded MCP/API calls described by the skill.
能力标签
cryptorequires-walletrequires-paid-servicerequires-sensitive-credentials
能力评估
Purpose & Capability
The stated purpose is explaining EVM/RPC/provider/wallet/account-abstraction errors and selector/catalog lookup; the artifact’s API, MCP, and package-import guidance matches that purpose.
Instruction Scope
Instructions are bounded around specific revert.wtf endpoints and MCP tools, and explicitly warn against broad catalog dumps and submitting private keys, seed phrases, access tokens, unreleased exploit details, or personal data.
Install Mechanism
The package contains only a single non-executable SKILL.md file, with no scripts, dependencies, or install-time actions.
Credentials
The skill involves crypto-related debugging and hosted API calls, so users should treat submitted error payloads as data shared with an external service; this is disclosed and purpose-aligned.
Persistence & Privilege
No persistence, privilege escalation, background worker, credential-store access, or local indexing behavior is present in the artifact.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install revertwtf-agent-api
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /revertwtf-agent-api 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v0.1.0
- Initial release of revertwtf-agent-api skill. - Provides safe use of revert.wtf via bounded HTTP APIs, MCP tools, and package subpaths. - Introduces MCP tools for explaining errors, selector resolution, catalog search, and Blockscout chain checks. - Documents safe usage patterns for hosted HTTP API endpoints. - Guides against large bundle imports in browser/client environments—prefer small helpers or API calls. - Includes precise package import recommendations for both server and browser contexts.
元数据
Slug revertwtf-agent-api
版本 0.1.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

revert.wtf Agent API Skill 是什么?

Use revert.wtf safely from agents, tools, browsers, or integrations through bounded HTTP APIs, MCP tools, and lightweight package subpaths without downloadin... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 18 次。

如何安装 revert.wtf Agent API Skill?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install revertwtf-agent-api」即可一键安装,无需额外配置。

revert.wtf Agent API Skill 是免费的吗?

是的,revert.wtf Agent API Skill 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

revert.wtf Agent API Skill 支持哪些平台?

revert.wtf Agent API Skill 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 revert.wtf Agent API Skill?

由 mrtdlgc(@mrtdlgc)开发并维护,当前版本 v0.1.0。

💬 留言讨论