{
  "markdown": "# HyperXosist Agent\r\n\r\n[![CI](https://github.com/KG-NINJA/HyperXosist-Agent/actions/workflows/ci.yml/badge.svg)](https://github.com/KG-NINJA/HyperXosist-Agent/actions/workflows/ci.yml)\r\n[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)\r\n[![Version](https://img.shields.io/badge/version-2.6.0-brightgreen.svg)](CHANGELOG.md)\r\n[![Live](https://img.shields.io/badge/demo-GitHub%20Pages-black.svg)](https://kg-ninja.github.io/HyperXosist-Agent/)\r\n\r\n**Agent-first toolkit and Remote MCP server for noise-reduced X (Twitter) search planning and Feedback-to-Fix engineering handoffs.** It generates structured search plans and official X search URLs for humans and AI agents. The planning and synthetic demonstration workflow does **not** directly scrape X. Optional **Grok Build** mode is default off.\r\n\r\n| | |\r\n|---|---|\r\n| **Live demo** | https://kg-ninja.github.io/HyperXosist-Agent/ |\r\n| **Repository** | https://github.com/KG-NINJA/HyperXosist-Agent |\r\n| **Agent entry** | https://kg-ninja.github.io/HyperXosist-Agent/llms.txt |\r\n| **CLI** | `npx hyperxosist plan \"…\" --json` |\r\n| **Version** | 2.6.0 |\r\n\r\n> 日本語の要点: X 公式検索用クエリを組み立てる静的ツールです。人間UIとAgentのplanning / filtering / handoffは無料。production executionのみx402支払い対象です。検索結果の埋め込みや自動投稿はしません。\r\n\r\n## Agent-readable summary\r\n\r\nHyperXosist-Agent helps an AI agent turn a research intent into a structured search plan, an official X search URL, supplied-feedback filtering results, and a Signal-to-Fix engineering handoff. It is a planning and handoff layer: it does not claim to scrape X directly, publish posts, or deploy code automatically.\r\n\r\nFree capabilities include MCP discovery, planning, filtering, structured handoff, and local dry-run workflows. Automated production use of generated search URLs and external collection use the existing x402 endpoint and return HTTP 402 until valid payment proof is supplied. Payment requirements are published at [`payment-options.json`](https://api.kgninja.dev/payment-options.json).\r\n\r\nAgents must follow the applicable user or operator payment policy. They must not request private keys, seed phrases, wallet secrets, or sensitive user data. If a wallet requires human approval, the agent should ask the operator to review and approve the transaction, then retry the same request with valid x402 payment proof.\r\n\r\n---\r\n\r\n## Why this exists\r\n\r\nRaw X advanced search is powerful but easy to get wrong (spam, engagement bait, overlong excludes, one-angle keyword guesses). HyperXosist gives you:\r\n\r\n1. **Humans** — a fast dark UI to compose operators, templates, noise filters, and open official search tabs.\r\n2. **AI agents** — a sticky loop: plan → score → pay → collect → refine → Keep-filter → Grok Build / Signal-to-Fix → receipt.\r\n\r\n```\r\nDiscover → Plan → Score gate → Pay (x402) → Collect → Self-heal → Keep-filter → Grok Build / Handoff → Remember\r\n```\r\n\r\n---\r\n\r\n## Features\r\n\r\n### For everyone\r\n- **No X API / OAuth** — opens `x.com/search` with a built query\r\n- **Zero build** — static HTML/CSS/JS; works on GitHub Pages\r\n- **Privacy-friendly human path** — form history in `localStorage` only\r\n- **Noise Reduction** — Low / Medium / High with **priority-capped** excludes (safe query length)\r\n- **Advanced operators** — `from` / `to` / `@` / OR groups / hashtags / `url:` / engagement floors / media & reply filters / raw fragment\r\n- **Research templates** & date presets (24h → 1y)\r\n- **Shareable state** — URL hash `#s=...`\r\n\r\n### For AI agents\r\n- **Official Model Context Protocol (MCP) Server** — expose query planning, signal filtering, and downstream handoff capabilities to Cursor, Claude Code, and other LLM assistants. Read [docs/MCP.md](docs/MCP.md).\r\n- **`dispatchToolCall` / `runTool`** — real multi-runtime tool dispatch (no hand-written mapping)\r\n- **`toOpenAITools()` / `toAnthropicTools()`** — drop-in schemas for GPT / Claude / Grok / Llama\r\n- **CLI** `bin/hyperxosist.js` — shell agents get `--json` plan / dispatch / keep / handoff\r\n- `exportKeepOnlyJson` — keep-only machine export for any coding agent\r\n- `planFromIntent` / multi-angle **missions**\r\n- `scoreQuery` before spending **$0.01** per paid call\r\n- `suggestRefinements` when results are empty or noisy\r\n- `buildSignalToFixPipeline` → full linked loop into [Signal-to-Fix](https://kg-ninja.github.io/Signal-to-Fix/) (humans: free UI steps; agents: x402)\r\n- `buildHandoffPackage` → Signal-to-Fix keep-only PR handoff package\r\n- Discovery: [`signal-to-fix-pipeline.json`](https://kg-ninja.github.io/HyperXosist-Agent/signal-to-fix-pipeline.json)\r\n- Dual **JSON + Markdown** outputs on core APIs (any LLM style)\r\n- `buildAgentPrompt` — model-agnostic one-small-change implementation prompt\r\n- Transparent noise catalog: `exportNoiseCatalog` / `noise.extraTerms`\r\n- Optional **Grok Build** mode: `createGrokBuildSession` / `buildGrokBuildPrompt` (default off)\r\n- `agent-tools.json` — OpenAI-compatible tools (portable to Claude/Grok/Llama runtimes)\r\n- `llms.txt` + `AGENTS.md` multi-LLM discovery docs\r\n\r\n---\r\n\r\n## Quick start (human)\r\n\r\n1. Open the [live demo](https://kg-ninja.github.io/HyperXosist-Agent/).\r\n2. Enter keywords (and optional OR group, users, dates, engagement).\r\n3. Optionally enable **Noise Reduction** and pick a research template.\r\n4. Click **最新で検索** (Latest) or **話題で検索** (Top) — or `Ctrl+Enter` / `Ctrl+Shift+Enter`.\r\n5. Copy query, search URL, or a shareable state link.\r\n6. **Signal-to-Fix 手動連携（無料）**: 投稿を Collected signals に貼る → **Handoff 生成** → **Signal-to-Fix 用をコピー** → [Signal-to-Fix](https://kg-ninja.github.io/Signal-to-Fix/) で Analyze → **keep のみ**使う。\r\n\r\nLocal UI:\r\n\r\n```bash\r\ngit clone https://github.com/KG-NINJA/HyperXosist-Agent.git\r\ncd HyperXosist-Agent\r\nnpm run serve\r\n# → http://localhost:5173\r\n```\r\n\r\n---\r\n\r\n## Quick start (AI agent)\r\n\r\n**Discovery order**\r\n\r\n1. [llms.txt](https://kg-ninja.github.io/HyperXosist-Agent/llms.txt)\r\n2. [AGENTS.md](https://kg-ninja.github.io/HyperXosist-Agent/AGENTS.md)\r\n3. [agent-use.json](https://kg-ninja.github.io/HyperXosist-Agent/agent-use.json)\r\n4. [agent-tools.json](https://kg-ninja.github.io/HyperXosist-Agent/agent-tools.json)\r\n5. [x402-payment.json](https://kg-ninja.github.io/HyperXosist-Agent/x402-payment.json)\r\n\r\n**Any runtime in 3 lines**\r\n\r\n```js\r\nconst HyperXosistAgent = require('hyperxosist-agent'); // or ./agent-api.js\r\n\r\n// 1) Register tools with your model runtime\r\nconst tools = HyperXosistAgent.toOpenAITools();     // or toAnthropicTools()\r\n\r\n// 2) When the model calls a tool — one dispatcher for all shapes\r\nconst { ok, result } = HyperXosistAgent.dispatchToolCall(name, args);\r\n```\r\n\r\n**Shell / CLI (no embed)**\r\n\r\n```bash\r\nnpm run cli -- plan \"Find product feedback about Acme for PR specs\" --json\r\n# or: node bin/hyperxosist.js dispatch hyperxosist_plan_from_intent \\\r\n#        --args '{\"intent\":\"Find feedback about Acme\"}' --json\r\n```\r\n\r\n**One call (library sticky loop)**\r\n\r\n```js\r\n// Node\r\nconst HyperXosistAgent = require('./agent-api.js');\r\n\r\nconst session = HyperXosistAgent.startAgentSession({\r\n  intent: 'Find product feedback about Acme for PR specs'\r\n});\r\n\r\nconst step = session.plan.primaryStep;\r\nif (step.score.recommendPay) {\r\n  const paid = step.paidRequest;\r\n  // POST paid.body → paid.endpoint\r\n  // expect HTTP 402 until x402 payment proof, then 200\r\n  // after authorization, open step.searchUrl and collect post texts\r\n}\r\n\r\nconst keepOnly = HyperXosistAgent.exportKeepOnlyJson(\r\n  ['...candidate posts...'],\r\n  { productName: 'Acme' }\r\n);\r\n// → keepOnly.texts / keepOnly.signalToFixInput / keepOnly.agentPrompt\r\n\r\nconst handoff = HyperXosistAgent.buildHandoffPackage({\r\n  productName: 'Acme',\r\n  feedback: ['...candidate posts...']\r\n});\r\n// → handoff.signalToFix.input into Signal-to-Fix (keep-only only)\r\n// → handoff.grokBuild.prompt for Grok Build\r\n\r\n// Grok Build path\r\nconst grok = HyperXosistAgent.createGrokBuildSession(\r\n  'Grok Build code improvement for Acme',\r\n  { product: 'Acme', targetArea: 'auth' }\r\n);\r\nconst prompt = HyperXosistAgent.buildGrokBuildPrompt({\r\n  productName: 'Acme',\r\n  targetArea: 'auth',\r\n  feedback: ['login button does nothing on Safari']\r\n});\r\n// → paste prompt.markdown into Grok Build\r\n```\r\n\r\n**CLI dry-run (no payment, no network required for planning)**\r\n\r\n```bash\r\nnpm test\r\nnpm run quickstart\r\n# or: node examples/quickstart.mjs \"Weekly monitor about MyProduct\"\r\n```\r\n\r\n### ChatGPT Site Tools / WebMCP\r\n\r\nThe GitHub Pages site feature-detects `document.modelContext.registerTool(...)` and exposes four Site Tools in compatible browser environments.\r\n\r\nFree, local, and read-only:\r\n\r\n- `hyperxosist_search_plan`\r\n- `hyperxosist_filter_signals`\r\n- `hyperxosist_build_handoff`\r\n\r\nPaid production boundary:\r\n\r\n- `hyperxosist_execute` — x402 v2 production execution through `https://api.kgninja.dev/hyperxosist-query`\r\n\r\nThe first call omits `paymentSignature` and returns the server's `PAYMENT-REQUIRED` requirements. A compatible wallet or facilitator authorizes the payment; the caller then retries with the opaque `PAYMENT-SIGNATURE` value and `confirmPayment: true`. The site does not create, request, store, log, or echo private keys, seed phrases, or wallet passwords.\r\n\r\nThe paid tool is marked consequential and non-idempotent. Clients must not auto-retry a confirmed payment call. Browsers without WebMCP support continue to use the normal human UI unchanged.\r\n\r\n### MCP: Local and Remote\r\n\r\nThe shared MCP core contains the same economic boundary as WebMCP:\r\n\r\n- Three free tools: planning, supplied-signal filtering, and engineering handoff.\r\n- One paid tool: `hyperxosist_execute`, which forwards an explicitly confirmed x402 request to the existing payment Worker.\r\n\r\n```bash\r\n# Local stdio for Cursor, Claude Code, and compatible clients\r\nnpm run mcp\r\n\r\n# Optional private Node Streamable HTTP server\r\nHYPERXOSIST_MCP_TOKEN=\"replace-me\" npm run mcp:remote\r\n```\r\n\r\nPublic production Remote MCP:\r\n\r\n- Endpoint: `https://mcp.kgninja.dev/mcp`\r\n- Health: `https://mcp.kgninja.dev/health`\r\n- Transport: Streamable HTTP\r\n- Public free-mode authentication: none\r\n- Private/self-hosted authentication: optional Bearer token\r\n- Currently deployed tools: `hyperxosist_search_plan`, `hyperxosist_filter_signals`, `hyperxosist_build_handoff`\r\n- `hyperxosist_execute`: implemented and tested in the repository, but requires an explicit production Cloudflare Worker deployment before the public Remote MCP advertises it\r\n\r\nMerging this repository does not deploy the Cloudflare Worker. Until that manual deployment occurs, the live Remote MCP remains the three-tool free service. The WebMCP paid tool is independent of that deployment and calls the already-live x402 endpoint directly.\r\n\r\nThe existing x402 route is the only verifier and settlement boundary. MCP and GitHub Pages never create or verify payment proofs. Remote MCP telemetry must not log request content or `PAYMENT-SIGNATURE`; settlement analytics remain authoritative in the x402 Worker.\r\n\r\n```bash\r\nnpm run openai:remote-check\r\nnpm run test:tool-selection\r\nnpm run test:paid-execution\r\nnpm run test:webmcp\r\n```\r\n\r\nSee [MCP setup and security](docs/MCP.md), [Remote MCP Worker deployment](workers/remote-mcp/README.md), and [ChatGPT App preparation](docs/CHATGPT_APP.md).\r\n\r\n### Agent handoff dry-run (offline — recommended first step)\r\n\r\nDemonstrates the full **local** path from search intent → keep filter → Signal-to-Fix handoff → coding-agent prompt **without network access**:\r\n\r\n```bash\r\nnpm run agent-handoff-dryrun -- HyperXosist-Agent\r\n# or: node examples/agent-handoff-dryrun.mjs \"HyperXosist-Agent\"\r\n```\r\n\r\nWhat it does (all offline):\r\n\r\n1. Builds an agent session / mission from intent  \r\n2. Prints mission ID, subject, primary query, and search URL  \r\n3. Uses **built-in sample feedback** (not live X data)  \r\n4. Runs `filterKeepSignals` (keep vs discard)  \r\n5. Builds `buildHandoffPackage`  \r\n6. Prints Signal-to-Fix input JSON preview  \r\n7. Prints the coding-agent implementation prompt (Markdown)  \r\n\r\n**Important:** this is a **local dry-run**. It does **not** scrape X, collect real posts, open the search URL, post anything, or perform x402 payment. Real agent production search still requires x402 after `scoreQuery`.\r\n\r\n### Payment policy (agents)\r\n\r\n| Use | Cost |\r\n|-----|------|\r\n| Human browser UI and manual official X URL use | Free |\r\n| Local planning, scoring, filtering, and handoff | Free |\r\n| WebMCP free tools | Free |\r\n| Remote MCP deployed planning/filtering/handoff tools | Free |\r\n| `hyperxosist_execute` production execution | **x402 paid**; current metadata states 0.01 USDC on Base |\r\n\r\nExecution flow:\r\n\r\n1. Call `hyperxosist_execute` with `input` and no payment signature.\r\n2. Receive HTTP 402 requirements, including the standard `PAYMENT-REQUIRED` response header.\r\n3. Authorize through a compatible x402 wallet or facilitator.\r\n4. Retry with `paymentSignature` and `confirmPayment: true`.\r\n5. Read the result and optional `PAYMENT-RESPONSE`; do not automatically repeat the confirmed call.\r\n\r\n`payment-options.json` is authoritative for current price, asset, network, payee, and facilitator data. GitHub Pages and MCP do not verify or settle payment.\r\n\r\n---\r\n\r\n\r\n\r\n## Missions agents re-run\r\n\r\n| ID | Purpose |\r\n|----|---------|\r\n| `product_feedback_radar` | Complaints / feature asks / bugs |\r\n| `signal_to_fix_pipeline` | Harvest → PR handoff loop |\r\n| `competitive_intel` | Mentions + switching language |\r\n| `weekly_monitor` | 7-day cron-friendly window |\r\n| `launch_pulse` | Launch / incident discourse |\r\n| `osint_entity` | from / mention / reply-to angles |\r\n| `grok_code_improvement_radar` | Grok Build: bugs / small asks / DX |\r\n| `ui_ux_feedback_harvest` | Frontend / UI friction for Grok |\r\n| `performance_complaint_detector` | Latency / jank for Grok |\r\n\r\nFull catalog: [missions.json](https://kg-ninja.github.io/HyperXosist-Agent/missions.json)\r\n\r\n---\r\n\r\n## API surface (v2.6)\r\n\r\n| Method | Role |\r\n|--------|------|\r\n| **`dispatchToolCall` / `runTool`** | **Execute any tool name** (OpenAI/Anthropic/plain shapes) |\r\n| **`toOpenAITools` / `toAnthropicTools`** | Drop-in tool schemas |\r\n| **`exportKeepOnlyJson`** | Keep-only JSON + S2F input + agent prompt |\r\n| `startAgentSession(opts?)` | Universal session (optional `mode:'grok'`) |\r\n| `planFromIntent(intent)` | NL → mission + scored paid steps + `.markdown` |\r\n| `buildMission(id, ctx)` | Named multi-angle campaign |\r\n| `scoreQuery(input)` | 0–100 + `recommendPay` + `.markdown` |\r\n| `suggestRefinements(input, signals)` | Self-heal + `.markdown` |\r\n| `buildHandoffPackage` | Signal-to-Fix + `agentPrompt` (any LLM) |\r\n| `buildAgentPrompt(opts)` | Universal one-small-change prompt |\r\n| `exportNoiseCatalog` / `customizeNoiseRules` | Transparent noise editing |\r\n| `filterKeepSignals` / `scoreTechnicalDepth` | Keep-only signal quality |\r\n| `buildGrokBuildPrompt` / `createGrokBuildSession` | **Optional** Grok mode |\r\n| `buildQuery` / `buildSearchUrl` / `buildShareUrl` | Query + shareable state |\r\n| `buildPaidRequest` / `buildBatch` | x402 payloads |\r\n| `getToolDefinitions` / `listMissions` | Catalogs (Grok tools opt-in; `format:'anthropic'`) |\r\n\r\n### CLI surface\r\n\r\n```bash\r\nnpx hyperxosist plan \"…\" --json\r\nnpx hyperxosist dispatch <toolName> --args '{…}' --json\r\nnpx hyperxosist tools --format openai|anthropic|full --json\r\nnpx hyperxosist keep --product X --feedback '[…]' --export-keep-only --json\r\nnpx hyperxosist handoff --product X --feedback '[…]' --json\r\nnpx hyperxosist pipeline --product X --json\r\n```\r\n\r\n---\r\n\r\n## Repository layout\r\n\r\n```\r\nindex.html, app.js, style.css, favicon.svg   # Human UI\r\nagent-api.js                                # Single-source API (Node + browser)\r\nbin/hyperxosist.js                          # Universal CLI (plan/dispatch/tools/keep…)\r\nagent-use.json                              # Agent manifest (sticky loop)\r\nagent-tools.json                            # OpenAI-compatible tools\r\nmissions.json                               # Mission catalog\r\nllms.txt, AGENTS.md                         # Agent discovery / playbook\r\nx402-payment.json                           # Payment metadata\r\ntop30_repost_blacklist.json                 # Bait phrase reference\r\nexamples/quickstart.mjs                     # CLI demo\r\nexamples/agent-handoff-dryrun.mjs           # Offline handoff dry-run (intent→keep→S2F)\r\nexamples/agent-session.example.json         # Session shape\r\npaid-execution.js                         # Shared x402 402→signature→200 bridge\r\nwebmcp.js                                 # Browser Site Tools adapter\r\naccess-policy.json                        # Machine-readable free/paid boundary\r\nmcp/tools.js, mcp/core.js                  # Shared MCP schemas + dispatch\r\nmcp/server.js, mcp/remote-server.js         # Local stdio + Remote Streamable HTTP\r\nexamples/openai-remote-mcp.mjs              # Responses API Remote MCP example\r\ntest/agent-api.test.js                      # Agent API tests\r\ntest/mcp-*.test.js                          # stdio, remote, security, schema tests\r\n.github/workflows/ci.yml                    # CI\r\nLICENSE, CHANGELOG.md, SECURITY.md, CONTRIBUTING.md\r\n```\r\n\r\n---\r\n\r\n## Deploy (GitHub Pages)\r\n\r\n1. Push to `main`.\r\n2. **Settings → Pages → Deploy from a branch → `main` / `/ (root)`**.\r\n3. Site: `https://<user>.github.io/HyperXosist-Agent/`\r\n\r\nNo build step. CI runs tests on every push/PR to `main`.\r\n\r\n### Public integration checks\r\n\r\n- [ ] UI loads on GitHub Pages  \r\n- [ ] `agent-use.json` / `x402-payment.json` / `access-policy.json` / `llms.txt` fetchable\r\n- [ ] Link to Signal-to-Fix `agent-use.json` works  \r\n- [ ] Unpaid `POST` to hyperxosist-query endpoint returns **402**\r\n\r\n---\r\n\r\n## Disclaimer\r\n\r\n- Not affiliated with X Corp.\r\n- This tool **only builds search queries / opens official search**. It does not embed timelines, auto-post, or replace the X API.\r\n- You are responsible for complying with X’s terms and applicable law when searching or collecting content.\r\n\r\n---\r\n\r\n## License\r\n\r\n[MIT](LICENSE) © 2026 KG-NINJA\r\n\r\n## Related\r\n\r\n- [Signal-to-Fix](https://kg-ninja.github.io/Signal-to-Fix/) — keep-only PR specs from feedback  \r\n- [CHANGELOG](CHANGELOG.md) · [CONTRIBUTING](CONTRIBUTING.md) · [SECURITY](SECURITY.md)\r\n\r\n## MCP discovery and catalog\r\n\r\nMachine-readable catalog metadata: https://kg-ninja.github.io/HyperXosist-Agent/mcp-catalog.json\r\n\r\n## For AI operators and agent builders\r\n\r\nUse HyperXosist-Agent when you need to discover customer complaints, product feedback, or pain points on X/Twitter without starting with noisy ad-hoc queries.\r\n\r\n**Common use cases:**\r\n\r\n- X/Twitter product-feedback discovery\r\n- Customer complaint and pain-point detection\r\n- Noise-reduced social listening query planning\r\n- Signal filtering for actionable feedback\r\n- AI-agent handoff generation for engineering teams\r\n\r\n**30-second Remote MCP quick start:**\r\n\r\n1. Connect an MCP client to https://mcp.kgninja.dev/mcp using Streamable HTTP.\r\n2. Call initialize, then tools/list.\r\n3. Start with hyperxosist_search_plan using a natural-language intent.\r\n4. Use hyperxosist_filter_signals and hyperxosist_build_handoff on collected text.\r\n5. For production execution, use `hyperxosist_execute`: inspect the unsigned 402 requirements, authorize with a compatible x402 wallet/facilitator, then retry once with explicit confirmation.\r\n\r\nFree MCP planning and handoff tools do not perform external collection. Human browser use remains free; automated production search execution is the paid boundary.\r\n\r\n## Optional signed artifact receipt\r\n\r\nWhen a downstream agent or audit workflow requires a service-signed record of an artifact digest check, use the opt-in [AVU buyer adapter](docs/AVU_BUYER.md). Handoff remains free; local-only checks make no network calls. The adapter checks availability, precheck digests, bound x402 terms and signed delivery. A buyer-owned wallet must explicitly authorize payment.\r\n\r\n```sh\r\nnpm run avu:demo    # offline synthetic example; never pays\r\nnpm run avu:status  # GET-only readiness check; never pays\r\n```\r\n\r\nThis is a source adapter, not a new deployed Remote MCP tool or an npm release. Live wallet/settlement compatibility still requires an authorized end-to-end test after service recovery.\r\n\r\n## Agent-to-agent service matching (source preview)\r\n\r\nUse `agent-matchmaker.mjs` or the new local stdio MCP server to match an agent's actual task, budget and deadline to five existing API services or an AVU artifact receipt. The client reuses the free public `/match` endpoint, checks advertised payment terms against live OpenAPI, and excludes local-only needs, unsupported capabilities and known same-operator purchases.\r\n\r\n```bash\r\nnpm run marketplace:demo                   # offline; no purchase\r\nnpm run marketplace:demo -- --live         # free discovery using synthetic demand\r\nnode mcp/matchmaker-server.mjs             # connect from an MCP-capable buyer\r\n```\r\n\r\nSee [the matching guide](docs/AGENT_MATCHMAKING.md) and [machine-readable entry](agent-marketplace.json). Six capabilities belong to one operator; external buyer demand, brokerage fees, real payment and revenue are not created by matching. The new tools are local source only and do not change the production Remote MCP deployment.\r\n",
  "bytes": 21185,
  "sha": "ca25e7221bac0b3edf2d7ce244a84e961225099a38bb04abd2c0949e3b761ea1",
  "repo_slug": "kg-ninja/hyperxosist-agent",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_kg_ninja_hyperxosist_agent_06bc33fa/readme"
}