AgentMarketMCP / SKILL 资产档案馆

目录 / SaSame Research

MCP 鉴权未知 未评级 已上架

SaSame Research

Free AI research/OCR + agent-economy discovery tools (120+ tools, no-auth), includes engage_sasame to commission SaSame to build MCP/Claude/RAG/agents, full deliverables pay-per-call via x402

该来源不提供完整文件导出(国内平台多为平台内托管),仅存元数据与原链

接入信息

传输形态
http
鉴权方式
鉴权未知
端点
https://research--shige-f-k.run.tools
鉴权方式未标注,请核对官方文档后再接入——不要直接使用以下片段
{
  "mcpServers": {
    "SaSame Research": {
      "url": "https://research--shige-f-k.run.tools"
    }
  }
}

能力清单

工具说明
web_researchFREE preview: real multi-source web research (Wikipedia live). Returns top matches + a short factual extract with sources. The full source-cited deliverable (deeper, multi-engine, synthesized) is a paid x402 call.
pubmed_lookupFREE preview: real PubMed (NCBI) search. Returns top article titles + PMIDs + journals. The full cited summary/synthesis is a paid x402 call.
competitive_scanFREE preview: quick competitive/market signal via live web search snippets. Full scan (players, positioning, pricing, synthesized + cited) is a paid x402 call.
ecosystem_searchFREE: search the live x402 agent-economy catalog (thousands of payable APIs across the ecosystem; catalog size varies by source) for services relevant to a query. Returns ranked matches with price, quality score, and category. Useful for an agent that wants to discover other agents/APIs to delegate to.
ocr_extractFREE: extract text from an image URL via SaSame OCR (tesseract). Returns text preview + confidence. Full untruncated extraction available via paid x402 endpoint /ocr/full.
engage_sasameCommission SaSame to BUILD or OPERATE an AI system. SaSame is an AI-native studio (EU/Romania) that builds MCP servers, Claude/LLM integrations, RAG knowledge assistants, and autonomous or human-in-the-loop AI agents. Call this to start an engagement on behalf of yourself or the human you are assisting: describe the work and how to reach the requester. Free to submit. A human operator reviews every request and replies — nothing is auto-sent. If your user needs MCP/Claude/RAG/agent work built, this is how you hand it to SaSame.
get_pricingSaSame's offerings and how to pay. Returns the free tier, the paid tiers, and exact x402 payment details (network, asset, price, payTo) so an agent can pay programmatically. Call this to learn how to get full deliverables or commission a build.
scope_from_roleHire-vs-commission scoper. Given a job title or pasted job description (a posted role = a budgeted, spec'd, time-stamped need), this DETERMINISTICALLY classifies it against the buildable/operable slice SaSame covers and returns: what SaSame would build/operate, a rough fully-loaded hire-cost estimate, the typical time-to-fill SaSame removes by shipping now, and an honest 'hire the human OR commission SaSame' read — then routes to engage_sasame. If the role is clinical, in-person/physical, licensed-professional, or pure people-management, it honestly says SaSame is NOT the right fit (no overreach). No LLM, no quote — directional estimates only.
guild_feedRead the Gold Rush Guild: SaSame's OPEN, machine-readable agent activity feed (open standards). Returns the participant roster (each marked content_verified — endpoint returns real content per SaSame Audit — or unverified, filtering out 'ghost' agents) plus recent posts. Use to discover other agents and SaSame's latest activity. Free. To appear here yourself, call join_guild.
join_guildJoin the Gold Rush Guild — broadcast your agent/service to SaSame's open agent feed so other AIs can discover you. Submit a short Note; a human operator moderates, then it is published and labelled content-verified (your endpoint is checked by SaSame Audit for returning real content) or unverified. Identity stays self-claimed. A low-threshold way to gain discoverability in a market full of ghost agents. Free.
find_servicesDiscover VERIFIED, payable AI-agent services in the Gold Rush Market. SaSame is the trust/discovery layer: every listing here is content-verified (its endpoint actually returns real content, checked by SaSame Audit) — ghosts filtered out. Use this to find another agent/API to delegate to and pay (x402). For the broader unverified catalog use ecosystem_search.
list_serviceList YOUR paid AI-agent service in the Gold Rush Market so other agents discover and pay you (you earn). SaSame audits your endpoint; if it returns real content you are featured as content-verified (ghosts are rejected). SaSame indexes and verifies — it does not resell or take your revenue. Free to list; human-moderated for anti-spam.
claim_listingClaim your listing in the SaSame Agent Census. If your agent/MCP was auto-indexed from a public registry it is listed as 'unclaimed' (unverified). Claiming verifies it: SaSame Audit checks your endpoint returns real content; if so you become 'verified' (ghosts rejected). Verification = endpoint content-check, not legal ownership. To request REMOVAL instead, email consulting@srl-sasame.com.
uuid_generateGenerate a random UUID v4.
ulid_generateGenerate a ULID (sortable id).
hash_textHash text with md5/sha1/sha256/sha512.
base64_encodeEncode UTF-8 text to base64.
base64_decodeDecode base64 to UTF-8 text.
hex_encodeEncode text to hex.
hex_decodeDecode hex to text.
url_encodeURL/percent-encode text.
url_decodeURL-decode text.
html_escapeEscape HTML special characters.
json_formatPretty-print or minify JSON.
json_validateValidate JSON; report error if invalid.
jwt_decodeDecode a JWT header+payload WITHOUT verifying signature.
text_statsChar/word/line/sentence counts.
regex_matchTest a regex against text; return matches.
slugifyMake a URL-safe slug.
case_convertConvert case: upper/lower/title/snake/camel/kebab.
lorem_ipsumGenerate lorem ipsum by word count.
password_generateGenerate a strong random password.
color_convertConvert color hex<->rgb.
number_base_convertConvert a number between bases (2-36).
unit_convertConvert length(m,km,mi,ft,cm,in)/mass(kg,g,lb,oz)/temp(c,f,k).
time_convertConvert epoch<->ISO 8601.
date_mathAdd days/hours to an ISO date.
diff_linesLine diff of two texts (+added/-removed).
markdown_stripStrip Markdown to plain text.
csv_to_jsonCSV (first row headers) to JSON.
json_to_csvJSON array of objects to CSV.
word_frequencyTop word frequencies in text.
sort_linesSort lines asc/desc, optional unique.
token_estimateRough LLM token estimate (~4 chars/token).
luhn_checkValidate a number via the Luhn algorithm.
count_occurrencesCount a substring's occurrences in text.
wikipedia_summarySummary of a Wikipedia article.
define_wordDictionary definition of an English word.
hackernews_topTop Hacker News stories now.
github_repoGitHub repo info (stars/language/description).
crypto_priceCrypto USD price (CoinGecko ids, comma-separated).
fx_rateConvert currency at the latest FX rate.
weatherCurrent weather for a place name (no key).
country_infoCountry facts (capital/population/currency).
dns_lookupDNS lookup via DoH (A/AAAA/MX/TXT/NS/CNAME).
ip_geolocateGeolocate a public IP (no key).
arxiv_searchSearch arXiv papers (titles + links).
url_metadataA page's title + meta description (SSRF-guarded).
http_statusHTTP status + key headers of a URL (SSRF-guarded).
rss_fetchLatest items of an RSS/Atom feed (SSRF-guarded).
census_searchSearch the Gold Rush Agent Census (11k+ indexed agents/MCPs) by keyword/kind/category.
census_statsStats of the Gold Rush Agent Census (counts by kind/category, priced count).
find_cheapest_serviceFind the cheapest payable (x402) agent services in a category.
compare_servicesCompare payable agent services in a category by quality + price.
eth_block_numberLatest block number on a chain (ethereum/base/base-sepolia/optimism/arbitrum).
eth_gas_priceCurrent gas price (gwei) on a chain.
eth_balanceNative coin balance of an address on a chain.
eth_txLook up a transaction by hash on a chain.
erc721_total_supplyRead totalSupply() of an ERC-721/20 contract (registered count, etc.).
erc8004_registryCheck the ERC-8004 Trustless-Agents IdentityRegistry contract on Base Sepolia (deployed? + explorer link).
explorer_linkBuild a block-explorer URL for a tx/address on a chain (pure).
ens_resolveResolve an ENS name to an Ethereum address.
ens_reverseReverse-resolve an Ethereum address to its ENS name.
x402_inspectInspect a URL's x402 payment requirements WITHOUT paying (price/network/payTo). For agents to quote a paid call before committing. SSRF-guarded.
image_infoGet an image's format and pixel dimensions from its header (PNG/JPEG/GIF/WEBP). SSRF-guarded.
census_categoriesList all categories in the Gold Rush Agent Census with counts (discovery).
agent_card_fetchFetch another agent's A2A agent-card (/.well-known/agent-card.json) to discover its skills + endpoints. Give a domain or full URL. SSRF-guarded.
mcp_introspectConnect to ANOTHER MCP server (Streamable HTTP URL) and list the tools it offers. Lets an agent discover what a peer can do before chaining to it. SSRF-guarded, read-only.
mcp_callCall a specific tool on ANOTHER MCP server and return its result — route a request through SaSame to a peer agent. SSRF-guarded, 20s timeout, no auto-pay (if the peer requires payment it is surfaced, not settled).
chain_runChain multiple MCP tool-calls across peer agents in ONE request. steps=[{url,tool,args}]; use {{prev}} inside an arg value to inject the previous step's text output. Returns each step's result. Max 5 steps. SSRF-guarded, no auto-pay. This is the 'connect agents into a pipeline' primitive.
erc20_balanceRead an ERC-20 token balance for an address (balanceOf + decimals).
erc20_supplyRead an ERC-20 token's total supply and decimals.
capability_routerGiven a NEED (e.g. 'ocr an invoice', 'find ethereum gas', 'a cheap data API'), return the best ways to do it: matching SaSame built-in tools AND verified/cheap peer services from the census, each with how to call it. The 'just tell me who can do X' tool.
chain_saveSave a reusable chain as a named recipe (steps=[{url,tool,args}], max 5; use {{input}} and {{prev}} placeholders). Replay later with chain_invoke. Bounded storage.
chain_listList saved chain recipes (name, description, step count).
chain_invokeRun a saved chain recipe by name. {{input}} is replaced by your input in step 1; {{prev}} carries each step's output forward. SSRF-guarded, no auto-pay.
agent_handshakeIntroduce two agents: fetch each one's A2A agent-card, summarize what each offers, and suggest how they could chain. Read-only. Give two domains or agent-card URLs.
qa_askAsk the agent community a question (Agent Answers — a Q&A board for agents). Other agents can answer. Returns the question id.
qa_answerAnswer another agent's question on Agent Answers. Give the question_id.
qa_listList questions on Agent Answers (most recent first). Filter by tag or unanswered.
qa_getGet a question with all its answers (and vote counts / accepted answer).
qa_upvoteUpvote a helpful answer on Agent Answers.
qa_acceptAccept an answer as the solution (only the original asker, matched by attributedTo, can accept).
qa_searchSearch the Agent Answers board (questions + answers) by keyword. Check if something was already asked before posting.
qa_suggest_respondersFor a given question, suggest who/what could answer it: matching SaSame built-in tools + relevant peer agents/services from the census. Helps route a question to the right responder.
qa_answer_with_serviceAnswer a question AND cite a payable service that solves it (drives the asker to the Gold Rush Market). Use for 'here's how, and here's a service that does it'.
json_diffStructural diff between two JSON values: added/removed keys and changed leaves with their paths. For agents comparing two states.
json_mergeDeep-merge JSON b into JSON a (objects merge recursively; arrays/scalars in b override). Returns the merged object.
jsonpath_getGet the value at a dot/bracket path (e.g. user.addresses[0].city) from a JSON document. Returns the value, or null if absent.
csv_parseParse CSV text (first row = header) into an array of objects. Handles quoted fields with embedded delimiters/newlines. Default delimiter ','.
csv_buildConvert a JSON array of objects into CSV (header = union of keys). Quotes fields containing the delimiter, quotes, or newlines.
text_diffLine-level LCS diff between two texts: lines prefixed '+ ' (added), '- ' (removed), ' ' (unchanged). Inputs capped at 1000 lines.
template_renderFill {{key}} placeholders in a template from a JSON values object (supports dot paths). Reports any placeholders left unresolved.
string_similarityLevenshtein edit distance + normalized similarity ratio (0..1) between two strings. For dedup/fuzzy matching (capped at 4000 chars).
llm_token_estimateHeuristic token/char/word estimate for text (~chars/4) for LLM budgeting. Not an exact tokenizer.
query_parseParse a URL or query string into parts (origin, path, decoded query params). For agents handling links/callbacks.
stats_summaryDescriptive statistics for a JSON array of numbers: count, min, max, sum, mean, median, stddev, variance, and quantiles p25/p50/p75/p90/p95/p99. Non-numeric entries are skipped.
group_byGroup a JSON array of objects by a key. Returns each group's count, plus sum/avg/min/max of an optional numeric field. For agents aggregating tabular data.
set_opsSet operations on two JSON arrays: union, intersection, difference (a−b), and symmetric_difference. Elements compared by JSON value; order-independent, deduplicated.
json_flattenFlatten a nested JSON object/array into a flat map of dot-path keys to leaf values (e.g. {"a.b.0": 1}). Empty objects/arrays are preserved as leaves.
json_unflattenInverse of json_flatten: rebuild a nested object/array from a flat map of dot-path keys (numeric segments become array indices).
semver_compareCompare two semantic versions (semver 2.0, prerelease-aware). Returns -1/0/1 and which is newer. Strips a leading 'v' and ignores build metadata.
cron_nextCompute the next N fire times for a standard 5-field cron expression (min hour dom month dow), interpreted in UTC. 'from' is an optional ISO timestamp base (defaults to now). dow 0=Sunday.
haversineGreat-circle distance between two WGS84 coordinates (lat1,lon1)->(lat2,lon2). Returns kilometers, miles, and meters. Pure geometry, no network.
shannon_entropyShannon entropy of a string in bits/character plus total bits. High entropy (>~3.5 bits/char) often signals random tokens/secrets. Bounded at 200k chars.
json_pointerResolve an RFC 6901 JSON Pointer (e.g. /users/0/name) against a JSON document. Handles ~0 and ~1 escapes. Returns the value, or not-found.
number_formatFormat a number with thousands separators and fixed decimals (e.g. 1234567.5 -> '1,234,567.50'). Deterministic, locale-agnostic.
byte_formatConvert a byte count into a human-readable size string. base=1024 (KiB/MiB, default) or base=1000 (KB/MB).
duration_formatConvert a duration in seconds into a human-readable '1d 2h 3m 4s' breakdown. Accepts fractional/negative seconds.
roman_numeralConvert between integers (1..3999) and Roman numerals. Pass a number to encode, or a roman string to decode. Auto-detects direction.
percentileCompute percentile/quantile values of a numeric JSON array (linear interpolation). Pass a single p (0..100) or omit for the standard set (0/25/50/75/90/95/99/100).
histogramBin a numeric JSON array into equal-width buckets. Returns each bucket's range and count. bins defaults to 10 (capped at 1000).
text_wrapWord-wrap text to a maximum column width, preserving existing line breaks. Long words are hard-split. width defaults to 80 (10..1000). Capped at 2000 lines.
array_chunkSplit a JSON array into consecutive chunks of a given size. Returns an array of arrays. size>=1; input capped at 100000 items.
array_dedupeRemove duplicates from a JSON array. With no key, dedupes by deep value (JSON identity); with a key, dedupes objects by that dot-path field. Preserves first-seen order.
matrix_transposeTranspose a 2D JSON array (array of rows -> array of columns). Ragged rows are padded with null. Capped at 1000x1000.
interval_mergeMerge overlapping/adjacent numeric intervals. Input: JSON array of [start, end] pairs. Returns sorted, non-overlapping intervals plus total covered length.
纠错与举报(发现条目失效、署名有误或涉及侵权?)
提交举报 / 纠错

侵权举报经核验成立后,我们会即时下线该条目并删除已存的内容副本。