跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
AI 与知识库

anysite-mcp

How to use the anysite MCP server effectively - the meta-tools (discover, execute, get_page, query_cache, export_data, search_requests, merge_data with join_on enrichment), the interactive entity table and lead review (show_entity_table, review_leads), the source map for GTM signals (funding, hiring, tech stack, reviews, news, launches), email finding cascades, domain->company resolution, and cost-aware calling patterns. Consult this before any anysite data work. Use when unsure which source or endpoint covers a data need, how much a call costs / how many credits, why an endpoint is 'not found', how to reuse a cache_key, how to paginate or re-filter cached results, or how to combine sources into a signal chain.

anysiteio · 已有静态检查记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add anysiteio/agent-skills --skill anysite-mcp

技能说明


name: anysite-mcp description: How to use the anysite MCP server effectively - the meta-tools (discover, execute, get_page, query_cache, export_data, search_requests, merge_data with join_on enrichment), the interactive entity table and lead review (show_entity_table, review_leads), the source map for GTM signals (funding, hiring, tech stack, reviews, news, launches), email finding cascades, domain->company resolution, and cost-aware calling patterns. Consult this before any anysite data work. Use when unsure which source or endpoint covers a data need, how much a call costs / how many credits, why an endpoint is 'not found', how to reuse a cache_key, how to paginate or re-filter cached results, or how to combine sources into a signal chain.

Anysite MCP — usage guide

The anysite MCP exposes hundreds of data sources through universal meta-tools, an interactive table with lead review, and the crm_* family (see Working with CRM). This skill is the map: how to call them, which sources cover which GTM need, and how to not waste credits.

The meta-tools

ToolPurposeCredits
discover(source, category)List endpoints + exact params for a source/categoryfree
execute(source, category, endpoint, params)Run an endpoint; returns first 10 items + cache_keypaid
get_page(cache_key, offset, limit)Page through a cached resultfree
query_cache(cache_key, conditions, sort_by, sort_order, aggregate, group_by, limit, offset)Filter/sort/aggregate cached data with SQL-like opsfree
export_data(cache_key, output_format, list_unpack)Export cached data — output_format json (default) / csv / jsonl; list_unpack = how many nested-array elements to expand into CSV columns (default 1)free
search_requests(source, category, endpoint, query, since, until, limit, offset)Find past execute() calls and their cache_keys — 7-day history, works across sessionsfree
merge_data(cache_keys, dedupe_by, join_on, combine, unmatched)Stack results (append + dedupe_by) or ENRICH the first key's rows with the later keys by join_on — see Tables, joins and lead reviewfree
show_entity_table(cache_key, title, initial_filters, sort, columns, group_by, inherit_state_from)Interactive table of companies/people for the user — only when the result carries view.type = "entity_table"free
review_leads(cache_key, title) / record_review(cache_key, row_ids, decision)One-company-at-a-time Yes/No/Skip cards over a companies table; record_review saves answers given in chatfree

Rules that prevent 90% of failures

  1. Always discover before execute. Endpoint names and params are not guessable, and a wrong source name returns the full source list — a wrong guess self-corrects for free. execute takes the endpoint NAME exactly as discover returns it (products_reviews), never a REST path segment (reviews) — resolution is an exact-match lookup.
  2. Never guess identifiers. LinkedIn aliases, URNs, Crunchbase aliases, Greenhouse board tokens are unpredictable. Resolve them through the search endpoint of the same source first.
  3. Re-use the cache — it outlives the session. execute returns a cache_key; further filtering, sorting, counting and paging of that result is free, and the cache lives for 7 days across sessions. Before any paid execute, check search_requests (free) for a recent identical call — same endpoint, matching params — and reuse its cache_key via query_cache/get_page instead of refetching (verified live: a two-day-old cache_key from another session served in full). Freshness rule: reuse when the data's age is fine for the task (enrichment firmographics — usually yes; "what's new today" — no).
  4. Cheap-first cascade. When several endpoints can answer, call the cached/DB one first (*/db/*, *sql* endpoints, ~1 credit) and the live one only for the remainder.
  5. Estimate volume before bulk runs — plan-aware. First know the user's plan (the CRM profile stores it after setup; if unknown, ask once: MCP Unlimited or credit-based?).
    • Credit-based plan: before anything above ~100 calls, state the estimate (N targets × credits-per-call) and get a nod. Prefer cheap DB endpoints, batch hard.
    • MCP Unlimited: credit warnings off, but keep batch sizes sane anyway — the real limits are latency and upstream rate limits, so cap sweeps the same way and say "this will take ~N minutes" instead of a price.
  6. Live LinkedIn search fails as an empty list, not an error. search_users and search_companies return {"results":[]} on queries that just don't hit ("stripe", "databar" both came back empty live, while "microsoft" worked) — it is not a broken key. On empty, switch to the search_sql_* DB endpoints; do NOT retry with broader keywords. (This is why reverse-lookup via live search_users is best-effort, not "usually one match".)
  7. gdelt is slow by design, not broken — 10–50s per call is normal (upstream per-IP throttling), and worst cases exceed the MCP client's silent-call timeout, which looks like a hang. Endpoints: gdelt/articles/articles_search and articles_context (timespan like 3d/1w or start_datetime YYYYMMDDHHMMSS; count ≤250). Keep it OUT of per-account sweep loops (use techmeme / google news — seconds); fine for a one-off deep media dive with a "takes a minute" warning.

Tables, joins and lead review

Lists of companies and people come back from execute() with view.type = "entity_table". In clients that render MCP Apps (claude.ai, Claude Desktop, ChatGPT) show them instead of pasting rows into chat; Claude Code renders no apps — there, work with get_page / query_cache / export_data.

  • Show, don't re-fetch. show_entity_table(cache_key, title=<the user's intent>). The user filters, sorts, selects, exports and as
阅读完整来源 ↗