"Python runner with streamable HTTP support" matching MCP tools:

• registerA

  agent-tools (x402 + MCP + A2A directory)A

  Self-register an x402 / MCP service in the agent-tools directory. Service owners and agents may submit new services here. Submissions are auto-reviewed instantly by x402 verification (no human gate): if the URL proves x402 payment support it is listed immediately and shows up in `search`; otherwise it is rejected or retried automatically. Listing is FREE. Dedup: if a service with the same canonical origin (scheme://host) already exists in the directory we return its slug instead of creating a duplicate submission. Same goes for a still-pending submission with the same origin. Rate limit: at most 5 pending submissions per client IP per 24h. Hits beyond that get `{error: rate_limited}` — try again later or email contact@agent-tools.cloud for bulk imports. Args: url: Public HTTPS URL of the service (the x402-payable endpoint or its homepage). Required. name: Human-friendly name. Defaults to the URL hostname. description: One-paragraph description (max ~2000 chars). mcp_url: If the service speaks MCP, its streamable-http endpoint. category: Free-form (e.g. "defi", "search", "social"). Use `list_categories` to align with existing taxonomy. chains: Networks the service accepts payment on (e.g. ["base", "solana"]). price_min_usdc: Lower bound of per-call price in USDC. price_max_usdc: Upper bound of per-call price in USDC. contact: Optional email / handle the directory team can reach you on for clarifications.

  ConnectorNo auth

• capabilitiesA

  MDEngineA

  Capability manifest of the hosted runners: LAMMPS version, installed packages, and every style by category with gpu=true (KOKKOS-accelerated) or gpu=false (exists, but runs on the pod's CPU cores at the GPU rate). Default = compact summary; runner=lammps&full=true returns the whole style table. Use preflight_deck to check a specific deck.

  ConnectorNo auth

• validate_pythonA

  CodeCraftA

  Validates a Python automation script that runs OUTSIDE the game, on three axes: Python syntax (using the real interpreter), Minecraft commands embedded in the script (against the official command index), and the shape of the /connect WebSocket message envelope. For behavior pack scripts use validate_script instead — Python does not run inside a pack. The embedded command check is the most valuable one: a command written from memory can look syntactically fine and still do nothing in the game. Only strings starting with / are treated as commands. If syntax could not be checked, syntaxChecked is false in the result; ok:true alone does not mean the syntax is valid.

  ConnectorNo auth

• kirk_bulk_howtoA

  Kirk — Unsupervised Structural Change DetectionA

  Return a self-contained stdlib Python client for scoring at ZERO per-call LLM tokens. Purpose: Hand the caller an HTTP consumer that runs locally so bulk scoring doesn't burn LLM tokens per book. Use when: You need to score more than ~200 books, or `kirk_score_book_batch` returned `batch_too_large`, or the caller is running an autonomous bulk workload that would otherwise pay per-tool-call LLM tokens for every book. Do not use when: You are running a one-off interactive call — a direct `kirk_score_book` invocation is simpler; don't route through the client for a single book. Capability class(es): Cost-steering / delivery-path tool. Hands the caller a runner that exercises the same C2 / C5 / C6 capabilities as the MCP scoring tools, but at zero per-call LLM token cost. Path fit: The returned client is an HTTP consumer of the same MCP endpoint. Production integrations run in-process under sealed-engine attestation — same binary sha as this endpoint. Contact Kavara for deployment options. Cost: 0 IU. Free tool. Once running locally, the returned client bills against the same tools it drives: single-book calls at 1 IU each, and batch calls at 1 IU per 50 books (minimum 1 IU per call). A full 500-book batch → 10 IU. No LLM tokens on top. Cost comparison (2.7M-book validation rerun via 500-book batches — ~5400 batches, 54000 IU billed either way): MCP via Sonnet 5: $1,968 LLM + $540 IU + ~15 days wall clock MCP via Haiku 4.5: $656 LLM + $540 IU + ~10 days Python client (this tool): $0 LLM + $540 IU + ~55 min Return structure: { "language": "python", "filename": "kirk_online_client.py", "requirements": str, "usage": str, "code": str (the client source, ~500 LOC), "example": str (2-line copy-paste demo) }

  ConnectorNo auth

• findagent_create_remote_mcpA

  FindAgentA

  List an EXTERNAL remote MCP server you run as a marketplace LISTING — for an MCP server hosted on YOUR OWN infrastructure that buyers connect their client straight to (FindAgent never proxies or runs it). Pass the listing basics (title/slug/tagline/description + example_prompts: 1–5 required; a listing is not categorised, so any category you send is ignored) and the remote endpoint as `server_url` (https) OR a parsed `server.json` object in `server_json`. The server's tools are auto-detected (a sandbox-gated live scan when available) — you can override with `tools` (name+description), `transport` (streamable-http|sse), and `auth_note` (what credential the server needs — NEVER a secret value). Creates a status=draft agent you own; then call findagent_submit_for_review IN THIS MCP CLIENT to submit it. The server URL is stored + displayed only; nothing executes on FindAgent. Before calling: findagent_check_slug + findagent_list_categories.

  ConnectorOAuth

• check_injectionA

  ContrastAPIA

  Scan source code for injection vulnerabilities: SQL injection, command injection, path traversal via unsafe string concatenation/unsanitized input. Supports Python, JavaScript, TypeScript, Java, Go, Ruby, Shell, Bash. Use to detect input-handling bugs; for secrets use check_secrets. Companion code-security tools: check_secrets (hard-coded credential detection), check_dependencies (known-CVE vulnerability audit), check_headers (live HTTP security-header validation), scan_headers (live HTTP scan via domain). Free: 30/hr, Pro: 500/hr. Returns {total, by_severity, findings}. No data stored.

  ConnectorNo auth

• submit_testA

  TestMyVibesA

  Queue a new TestMyVibes job for a given URL. You explicitly choose the runner: AI agent (headless Chromium + GPT-4o vision, fastest, deterministic for well-specified goals) or human checker (slower, better for visual/UX judgment calls). Returns a jobId you can poll with get_test_status.

  ConnectorNo auth

• submit_test_batchA

  TestMyVibesA

  Queue up to 20 AI tests at once and run them in parallel instead of one-after-another. Each test in the batch costs 1.15× its base credits (the parallel premium). Returns the shared batchId and a per-test breakdown so you can poll each jobId individually. Use this when you have an independent set of tests to run (e.g. signup + login + dashboard + settings + delete across one customer site) and want them done in minutes rather than queued through a serial worker. AI runner only — human-runner batching ships separately.

  ConnectorNo auth

• recommend_panelA

  Meta CouncilB

  Recommend the best expert panel for a query (semantic match with keyword fallback). Returns the top panel + confidence and the runner-up options — feed the result into run_council's panel argument. Requires authentication because the query may be sent to the configured embedding provider.

  ConnectorNo auth

• list_examplesA

  EpsilonA

  List runnable Epsilon integration examples (Python cron DCA, LangChain tools, Vercel AI SDK tools, Telegram bot). Fetch full source with get_example. No API key required.

  ConnectorAPI key

• compare_mcpsA

  mcp-judgeA

  Choosing between several MCP servers that do similar things? This tool decides for you. PAID $0.10 via x402 (USDC micropayment over HTTP 402 — no account or API key needed; on your first call without payment you receive the exact payment requirements, then retry with the X-PAYMENT header). Compares 2-5 MCP servers head-to-head with identical objective checks (handshake, tools/list, documentation coverage, latency, and a safe functional probe — paid tools are never called) and returns: a ranked list, the recommended winner, a plain-language explanation of why it won, and each server's full report — cheaper than separate evaluate_mcp calls on 3+ servers. Objective checks only — no human and no LLM opinion; it does not judge the real-world usefulness or safety of the content. Set 'urls' (required) to an array of 2-5 MCP endpoints (Streamable HTTP), e.g. ["https://a/mcp","https://b/mcp"].

  ConnectorNo auth

• gluecron_revoke_runnerA

  GluecronB

  Revoke a CI runner by id: nothing is dispatched to it from now on, its next heartbeat is refused and it stops accepting jobs. Cannot be undone — register a new runner instead. Returns {id, revoked}.

  Connector

  Destructive

  No auth

• get_cycleA

  EndoflifeA

  Get a single release cycle's support details for a product — release date, EOL, active-support end, latest patch, LTS, and any extended-support window. Use for a precise version question like "when does Python 3.9 lose support?". `product` is a slug from list_products; `cycle` is a version like "3.12", "20.04", "18". Keyless.

  ConnectorNo auth

• search_docsA

  Devdocs IoB

  Find which documentation SETS exist whose NAME matches a substring (e.g. "python" → Python 3.x, "react" → React). Returns doc SETS, NOT their content — this does NOT look up a function/method/API name. To search inside a doc for an entry like "Array.map" or "fetch", use search_index (slug + query).

  ConnectorNo auth

• test_python_package_runtimeA

  Python Package Runtime TestA

  Test a specific Python package/version on a supported Python/Linux target runtime using empirical execution. Returns scoped install evidence, usable-import evidence when requested, and a scoped result/receipt. Supported boundary: public PyPI on Linux x86_64 / CPython 3.13. This is not generic package advice, security verification, arbitrary Python execution, broad compatibility testing, feature behavior testing, or application integration testing.

  ConnectorNo auth

• assess_finish_time_goalA

  CourseProfilerA

  Assess a positive elapsed goal including saved dwell using the existing nine intensity/shape combinations (5% shapes). Preserve reference-plan movement locks, stops, checkpoint windows, cutoffs and saved GAP settings. Requires current v3 course and runner artifact IDs. Defaults to load_response_evidence_v2 and allow_profile_preparation=false; experimental v3 and in-memory preparation require explicit user approval, never silently retry with permission. Returns canonical support/classifications and bounded diagnostics plus a private complete planning_report artifact, not an applied/exportable plan. No source replacement, evidence selection or Strava access. Budget confidence is not certainty of finishing. Use get_artifact with report_artifact_id for full diagnostics; preserve omission counts and preparation provenance.

  ConnectorNo auth

• get_curl_retry_diagnosticA

  retry-traceA

  Get a downloadable localhost diagnostic for installed curl: compare default behavior with --retry 1 against HTTP 429/503 and Retry-After seconds/date headers. Returns the source URL, commands, requirements and report fields. This call does not run code. After source review, the caller may run it using Python 3.9+ and curl; execution uses loopback only and never uploads or publishes results. Useful when a local fixture can answer the task without a remote diagnostic run.

  ConnectorNo auth

• pit_supportA

  The PitA

  Optional ways to tip the Pit (card link, Bitcoin address, USDC on Base, and x402 tip URLs an agent can pay over HTTP) with network notes. Same as GET /v1/support. No key needed. Nothing is paid through this tool. Tips buy nothing and change nothing: no ranks, perks, visibility or priority.

  ConnectorAPI key

• evaluate_mcpA

  mcp-judgeA

  Call this BEFORE integrating any third-party MCP server: an unreliable or undocumented MCP wastes your tokens and can break your workflow. PAID $0.05 via x402 (USDC micropayment over HTTP 402 — no account or API key needed; on your first call without payment you receive the exact payment requirements, then retry with the X-PAYMENT header). Returns a complete objective quality report on the target: protocol conformance ('initialize' handshake), tool discovery via 'tools/list', per-tool and per-parameter documentation coverage, latency, and a safe functional probe — free tools are actually called, paid tools are only verified to be cleanly declared (MCP Judge never pays). You get a 0-100 score, A-F letter grade, per-criterion breakdown, and a plain-language summary you can act on. Tip: run the FREE quick_check first to get just the grade, and list_criteria to see exactly how scoring works. Objective checks only — no human and no LLM opinion; it does not judge the real-world usefulness or safety of the content. Set 'url' (required) to the target's MCP endpoint (Streamable HTTP), e.g. https://host/mcp.

  ConnectorNo auth

• attach_connectorA

  ValetA

  Attach an organization connector to a website or agent so it can call the connector's tools with the credential Valet holds. Attaching is a grant, and it is wider than it looks: for a website, every person who can open the page can call every tool the connector exposes. On a private website that is every member of the organization; on a password-protected or shared one it is everyone holding the password or the link. For an agent, the agent can. Valet does not narrow the connector's reach, so attach only what the thing needs, and check a website's access mode before you do. A website can hold only an organization connector that is an HTTP MCP server, on the sse or streamable-http transport; list_connectors with name marks which ones. A page calls the connector by its own name, which list_connector_tools reports. An app cannot hold a connector yet and the call says so. A connector that belongs to a single agent cannot be attached to anything else. Attaching a connector that is already attached changes nothing and is safe to repeat. Requires connecting a Valet account.

  ConnectorNo auth

• arena_playbookA

  LoomDeskA

  Returns the fee-farm playbooks the LoomDesk house agents run in the arena, as data and as a prompt: which tokens pass the farm bar (from market_tokens), how each plan ranks them, the band and size to open with arena_open, and the exact autopilot rules for arena_edit (kind auto), plus the one-slot runner rule. Use to follow a proven plan instead of guessing; a small model following it does what the house does. Not for real positions (plan_build). Returns: plans[] (key, title, way, band, usd, positions, rank, autopilot), farmBar, runnerBar, farmRules, runnerRules, prompt (the plan asked for, as text to follow), runnerPrompt. Behavior: read-only, free; the same words are MCP prompts loomdesk-farm and loomdesk-runner. Errors: none; an unknown plan returns the first.

  ConnectorNo auth