HyperXosist Agent

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.

Live demo	https://kg-ninja.github.io/HyperXosist-Agent/
Repository	https://github.com/KG-NINJA/HyperXosist-Agent
Agent entry	https://kg-ninja.github.io/HyperXosist-Agent/llms.txt
CLI	npx hyperxosist plan "…" --json
Version	2.6.0

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。

Agent-readable summary

HyperXosist-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.

Free 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.

Agents 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.

---

Why this exists

Raw X advanced search is powerful but easy to get wrong (spam, engagement bait, overlong excludes, one-angle keyword guesses). HyperXosist gives you:

1. Humans — a fast dark UI to compose operators, templates, noise filters, and open official search tabs.
2. AI agents — a sticky loop: plan → score → pay → collect → refine → Keep-filter → Grok Build / Signal-to-Fix → receipt.

Discover → Plan → Score gate → Pay (x402) → Collect → Self-heal → Keep-filter → Grok Build / Handoff → Remember

---

Features

For everyone

• No X API / OAuth — opens x.com/search with a built query
• Zero build — static HTML/CSS/JS; works on GitHub Pages
• Privacy-friendly human path — form history in localStorage only
• Noise Reduction — Low / Medium / High with priority-capped excludes (safe query length)
• Advanced operators — from / to / @ / OR groups / hashtags / url: / engagement floors / media & reply filters / raw fragment
• Research templates & date presets (24h → 1y)
• Shareable state — URL hash #s=...

For AI agents

• 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.
• dispatchToolCall / runTool — real multi-runtime tool dispatch (no hand-written mapping)
• toOpenAITools() / toAnthropicTools() — drop-in schemas for GPT / Claude / Grok / Llama
• CLI bin/hyperxosist.js — shell agents get --json plan / dispatch / keep / handoff
• exportKeepOnlyJson — keep-only machine export for any coding agent
• planFromIntent / multi-angle missions
• scoreQuery before spending $0.01 per paid call
• suggestRefinements when results are empty or noisy
• buildSignalToFixPipeline → full linked loop into Signal-to-Fix (humans: free UI steps; agents: x402)
• buildHandoffPackage → Signal-to-Fix keep-only PR handoff package
• Discovery: signal-to-fix-pipeline.json
• Dual JSON + Markdown outputs on core APIs (any LLM style)
• buildAgentPrompt — model-agnostic one-small-change implementation prompt
• Transparent noise catalog: exportNoiseCatalog / noise.extraTerms
• Optional Grok Build mode: createGrokBuildSession / buildGrokBuildPrompt (default off)
• agent-tools.json — OpenAI-compatible tools (portable to Claude/Grok/Llama runtimes)
• llms.txt + AGENTS.md multi-LLM discovery docs

---

Quick start (human)

1. Open the live demo.
2. Enter keywords (and optional OR group, users, dates, engagement).
3. Optionally enable Noise Reduction and pick a research template.
4. Click 最新で検索 (Latest) or 話題で検索 (Top) — or Ctrl+Enter / Ctrl+Shift+Enter.
5. Copy query, search URL, or a shareable state link.
6. 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 で Analyze → keep のみ使う。

Local UI:

git clone https://github.com/KG-NINJA/HyperXosist-Agent.git
cd HyperXosist-Agent
npm run serve
# → http://localhost:5173

---

Quick start (AI agent)

Discovery order

1. llms.txt
2. AGENTS.md
3. agent-use.json
4. agent-tools.json
5. x402-payment.json

Any runtime in 3 lines

const HyperXosistAgent = require('hyperxosist-agent'); // or ./agent-api.js

// 1) Register tools with your model runtime
const tools = HyperXosistAgent.toOpenAITools();     // or toAnthropicTools()

// 2) When the model calls a tool — one dispatcher for all shapes
const { ok, result } = HyperXosistAgent.dispatchToolCall(name, args);

Shell / CLI (no embed)

npm run cli -- plan "Find product feedback about Acme for PR specs" --json
# or: node bin/hyperxosist.js dispatch hyperxosist_plan_from_intent \
#        --args '{"intent":"Find feedback about Acme"}' --json

One call (library sticky loop)

// Node
const HyperXosistAgent = require('./agent-api.js');

const session = HyperXosistAgent.startAgentSession({
  intent: 'Find product feedback about Acme for PR specs'
});

const step = session.plan.primaryStep;
if (step.score.recommendPay) {
  const paid = step.paidRequest;
  // POST paid.body → paid.endpoint
  // expect HTTP 402 until x402 payment proof, then 200
  // after authorization, open step.searchUrl and collect post texts
}

const keepOnly = HyperXosistAgent.exportKeepOnlyJson(
  ['...candidate posts...'],
  { productName: 'Acme' }
);
// → keepOnly.texts / keepOnly.signalToFixInput / keepOnly.agentPrompt

const handoff = HyperXosistAgent.buildHandoffPackage({
  productName: 'Acme',
  feedback: ['...candidate posts...']
});
// → handoff.signalToFix.input into Signal-to-Fix (keep-only only)
// → handoff.grokBuild.prompt for Grok Build

// Grok Build path
const grok = HyperXosistAgent.createGrokBuildSession(
  'Grok Build code improvement for Acme',
  { product: 'Acme', targetArea: 'auth' }
);
const prompt = HyperXosistAgent.buildGrokBuildPrompt({
  productName: 'Acme',
  targetArea: 'auth',
  feedback: ['login button does nothing on Safari']
});
// → paste prompt.markdown into Grok Build

CLI dry-run (no payment, no network required for planning)

npm test
npm run quickstart
# or: node examples/quickstart.mjs "Weekly monitor about MyProduct"

ChatGPT Site Tools / WebMCP

The GitHub Pages site feature-detects document.modelContext.registerTool(...) and exposes four Site Tools in compatible browser environments.

Free, local, and read-only:

• hyperxosist_search_plan
• hyperxosist_filter_signals
• hyperxosist_build_handoff

Paid production boundary:

• hyperxosist_execute — x402 v2 production execution through https://api.kgninja.dev/hyperxosist-query

The 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.

The 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.

MCP: Local and Remote

The shared MCP core contains the same economic boundary as WebMCP:

• Three free tools: planning, supplied-signal filtering, and engineering handoff.
• One paid tool: hyperxosist_execute, which forwards an explicitly confirmed x402 request to the existing payment Worker.

# Local stdio for Cursor, Claude Code, and compatible clients
npm run mcp

# Optional private Node Streamable HTTP server
HYPERXOSIST_MCP_TOKEN="replace-me" npm run mcp:remote

Public production Remote MCP:

• Endpoint: https://mcp.kgninja.dev/mcp
• Health: https://mcp.kgninja.dev/health
• Transport: Streamable HTTP
• Public free-mode authentication: none
• Private/self-hosted authentication: optional Bearer token
• Currently deployed tools: hyperxosist_search_plan, hyperxosist_filter_signals, hyperxosist_build_handoff
• hyperxosist_execute: implemented and tested in the repository, but requires an explicit production Cloudflare Worker deployment before the public Remote MCP advertises it

Merging 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.

The 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.

npm run openai:remote-check
npm run test:tool-selection
npm run test:paid-execution
npm run test:webmcp

See MCP setup and security, Remote MCP Worker deployment, and ChatGPT App preparation.

Agent handoff dry-run (offline — recommended first step)

Demonstrates the full local path from search intent → keep filter → Signal-to-Fix handoff → coding-agent prompt without network access:

npm run agent-handoff-dryrun -- HyperXosist-Agent
# or: node examples/agent-handoff-dryrun.mjs "HyperXosist-Agent"

What it does (all offline):

1. Builds an agent session / mission from intent
2. Prints mission ID, subject, primary query, and search URL
3. Uses built-in sample feedback (not live X data)
4. Runs filterKeepSignals (keep vs discard)
5. Builds buildHandoffPackage
6. Prints Signal-to-Fix input JSON preview
7. Prints the coding-agent implementation prompt (Markdown)

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.

Payment policy (agents)

Use	Cost
Human browser UI and manual official X URL use	Free
Local planning, scoring, filtering, and handoff	Free
WebMCP free tools	Free
Remote MCP deployed planning/filtering/handoff tools	Free
hyperxosist_execute production execution	x402 paid; current metadata states 0.01 USDC on Base

Execution flow:

1. Call hyperxosist_execute with input and no payment signature.
2. Receive HTTP 402 requirements, including the standard PAYMENT-REQUIRED response header.
3. Authorize through a compatible x402 wallet or facilitator.
4. Retry with paymentSignature and confirmPayment: true.
5. Read the result and optional PAYMENT-RESPONSE; do not automatically repeat the confirmed call.

payment-options.json is authoritative for current price, asset, network, payee, and facilitator data. GitHub Pages and MCP do not verify or settle payment.

---

Missions agents re-run

ID	Purpose
product_feedback_radar	Complaints / feature asks / bugs
signal_to_fix_pipeline	Harvest → PR handoff loop
competitive_intel	Mentions + switching language
weekly_monitor	7-day cron-friendly window
launch_pulse	Launch / incident discourse
osint_entity	from / mention / reply-to angles
grok_code_improvement_radar	Grok Build: bugs / small asks / DX
ui_ux_feedback_harvest	Frontend / UI friction for Grok
performance_complaint_detector	Latency / jank for Grok

Full catalog: missions.json

---

API surface (v2.6)

Method	Role
dispatchToolCall / runTool	Execute any tool name (OpenAI/Anthropic/plain shapes)
toOpenAITools / toAnthropicTools	Drop-in tool schemas
exportKeepOnlyJson	Keep-only JSON + S2F input + agent prompt
startAgentSession(opts?)	Universal session (optional mode:'grok')
planFromIntent(intent)	NL → mission + scored paid steps + .markdown
buildMission(id, ctx)	Named multi-angle campaign
scoreQuery(input)	0–100 + recommendPay + .markdown
suggestRefinements(input, signals)	Self-heal + .markdown
buildHandoffPackage	Signal-to-Fix + agentPrompt (any LLM)
buildAgentPrompt(opts)	Universal one-small-change prompt
exportNoiseCatalog / customizeNoiseRules	Transparent noise editing
filterKeepSignals / scoreTechnicalDepth	Keep-only signal quality
buildGrokBuildPrompt / createGrokBuildSession	Optional Grok mode
buildQuery / buildSearchUrl / buildShareUrl	Query + shareable state
buildPaidRequest / buildBatch	x402 payloads
getToolDefinitions / listMissions	Catalogs (Grok tools opt-in; format:'anthropic')

CLI surface

npx hyperxosist plan "…" --json
npx hyperxosist dispatch <toolName> --args '{…}' --json
npx hyperxosist tools --format openai|anthropic|full --json
npx hyperxosist keep --product X --feedback '[…]' --export-keep-only --json
npx hyperxosist handoff --product X --feedback '[…]' --json
npx hyperxosist pipeline --product X --json

---

Repository layout

index.html, app.js, style.css, favicon.svg   # Human UI
agent-api.js                                # Single-source API (Node + browser)
bin/hyperxosist.js                          # Universal CLI (plan/dispatch/tools/keep…)
agent-use.json                              # Agent manifest (sticky loop)
agent-tools.json                            # OpenAI-compatible tools
missions.json                               # Mission catalog
llms.txt, AGENTS.md                         # Agent discovery / playbook
x402-payment.json                           # Payment metadata
top30_repost_blacklist.json                 # Bait phrase reference
examples/quickstart.mjs                     # CLI demo
examples/agent-handoff-dryrun.mjs           # Offline handoff dry-run (intent→keep→S2F)
examples/agent-session.example.json         # Session shape
paid-execution.js                         # Shared x402 402→signature→200 bridge
webmcp.js                                 # Browser Site Tools adapter
access-policy.json                        # Machine-readable free/paid boundary
mcp/tools.js, mcp/core.js                  # Shared MCP schemas + dispatch
mcp/server.js, mcp/remote-server.js         # Local stdio + Remote Streamable HTTP
examples/openai-remote-mcp.mjs              # Responses API Remote MCP example
test/agent-api.test.js                      # Agent API tests
test/mcp-*.test.js                          # stdio, remote, security, schema tests
.github/workflows/ci.yml                    # CI
LICENSE, CHANGELOG.md, SECURITY.md, CONTRIBUTING.md

---

Deploy (GitHub Pages)

1. Push to main.
2. Settings → Pages → Deploy from a branch → main / / (root).
3. Site: https://<user>.github.io/HyperXosist-Agent/

No build step. CI runs tests on every push/PR to main.

Public integration checks

• UI loads on GitHub Pages
• agent-use.json / x402-payment.json / access-policy.json / llms.txt fetchable
• Link to Signal-to-Fix agent-use.json works
• Unpaid POST to hyperxosist-query endpoint returns 402

---

Disclaimer

• Not affiliated with X Corp.
• This tool only builds search queries / opens official search. It does not embed timelines, auto-post, or replace the X API.
• You are responsible for complying with X’s terms and applicable law when searching or collecting content.

---

License

MIT © 2026 KG-NINJA

Related

• Signal-to-Fix — keep-only PR specs from feedback
• CHANGELOG · CONTRIBUTING · SECURITY

MCP discovery and catalog

Machine-readable catalog metadata: https://kg-ninja.github.io/HyperXosist-Agent/mcp-catalog.json

For AI operators and agent builders

Use HyperXosist-Agent when you need to discover customer complaints, product feedback, or pain points on X/Twitter without starting with noisy ad-hoc queries.

Common use cases:

• X/Twitter product-feedback discovery
• Customer complaint and pain-point detection
• Noise-reduced social listening query planning
• Signal filtering for actionable feedback
• AI-agent handoff generation for engineering teams

30-second Remote MCP quick start:

1. Connect an MCP client to https://mcp.kgninja.dev/mcp using Streamable HTTP.
2. Call initialize, then tools/list.
3. Start with hyperxosist_search_plan using a natural-language intent.
4. Use hyperxosist_filter_signals and hyperxosist_build_handoff on collected text.
5. For production execution, use hyperxosist_execute: inspect the unsigned 402 requirements, authorize with a compatible x402 wallet/facilitator, then retry once with explicit confirmation.

Free MCP planning and handoff tools do not perform external collection. Human browser use remains free; automated production search execution is the paid boundary.

Optional signed artifact receipt

When a downstream agent or audit workflow requires a service-signed record of an artifact digest check, use the opt-in AVU buyer adapter. 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.

npm run avu:demo    # offline synthetic example; never pays
npm run avu:status  # GET-only readiness check; never pays

This 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.

Agent-to-agent service matching (source preview)

Use 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.

npm run marketplace:demo                   # offline; no purchase
npm run marketplace:demo -- --live         # free discovery using synthetic demand
node mcp/matchmaker-server.mjs             # connect from an MCP-capable buyer

See the matching guide and machine-readable entry. 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.