Verified research with free-search
Use search results to find pages, and take facts only from the pages. A snippet
is an engine's summary of a page as it looked when it was crawled: it drops
qualifiers, merges editions, and is often years old. Use this workflow whenever
someone will rely on the answer.
Tool names below are the bare names on the search MCP server; your host may
show them with a prefix.
0. Fix today's date
Know today's date before judging anything as current. Take it from your system
context. Failing that, the retrieved … stamp on a fresh (uncached) search
result is the server's clock. "Latest", "this year" and "upcoming" are relative
to it. Do not infer the current year from search results or from memory.
1. Locate: search is for URLs only
search(query) finds candidate pages. Add freshness="month" (or
"day"/"week"/"year") when recency matters, include_domains=[...] once
you know the official domain, category= for papers, filings, news, datasets.
- Some facts have a registry, and the search asks it on its own when the
question names the thing plainly: a package ("latest fastapi version"), a
CVE id, two currencies ("100 usd to cny"), a place and a weather word, a
country and an indicator ("china gdp"), a city and a time word, a domain,
a coin, an iOS app. The record comes back as the first result with the
publisher's date, marked "Record sources", and can be cited as it is.
category= ("software", "security", "weather", "finance.fx", "stats",
"calendar", "reference") asks the same sources explicitly.
- Pick the primary source: the organizer's, vendor's, regulator's or author's
own page, the filing, the paper, the repository. Aggregators, listicles,
reposts, forums and AI-written summaries are useful only for finding it.
- Never copy a date, amount, name, rule or number from a snippet or the
Lead:
line into your answer.
- Search in the language of the source. A Chinese query is routed to a Chinese
index automatically; to force more of them use
engines=["so360", "quark", "baidu"].
2. Read the source itself
| Need |
Call |
| One page |
fetch(url) |
| Several candidates |
fetch_batch(urls) |
| Find and read in one round trip |
research(question, depth=3, include_domains=[...]) |
| Long PDF or DOCX (rules, notices) |
read_doc(source, start, length), next page at start + returned_chars |
| Dates, prices, event metadata as fields |
extract_structured(url): JSON-LD startDate, price, datePublished, dateModified |
| One question across 2-5 pages |
compare(question, urls) |
| A paper you are about to cite |
paper_graph(paper) for retraction and correction notices |
Divisions, regions, fee tables, prize breakdowns and deadlines usually sit on a
sub-page or in a linked PDF (rules, FAQ, 章程, 通知). If the landing page lacks
them, follow its links with fetch or read_doc before you go back to search.
If fetch returns a near-empty body, retry with render="browser".
3. Check that it is current
For every source you rely on, establish three things.
- When it was published or last updated. Look for
published <date> in a
fetch header, datePublished/dateModified from extract_structured, or
a date in the body. no publication date found means the page is undated.
Say so, and do not assume it is recent.
- When this copy was retrieved. Pages and searches are cached for up to 7
days, and
cached N days ago in the output means you are reading a stored
copy. For anything that changes (deadlines, prices, availability, standings,
versions, breaking news) pull it again with fetch(url, force_refresh=True),
or max_age_hours=0 on search and research.
- Which edition it describes. The year, season, version or round on the page
must match the one asked about. Last year's page for an annual event is the
most common wrong answer, because it is official and detailed and so looks
right. If only an older edition is published, report it as that edition and
say the current one is not announced.
freshness= is best-effort and keeps undated results, so a result that
survives freshness="month" may still be old. The search says how many results
it could date (dated: 3/10) and marks each one.
4. Corroborate
- A fact the user will act on needs one primary source, or two independent
ones. Independent means different origin: two outlets reprinting one press
release or wire story count as one source. Identical wording usually means a
shared origin.
- Prefer primary over secondary, and the newer of two primary pages.
- When sources disagree, do not average and do not pick silently. Report both
values, who says which, their dates, and which you treat as authoritative
and why.
5. Read the warnings
dated <date> on a result means the date came from a feed or an API field.
<date> (from snippet text) means it was lifted out of prose and may be a
date the page mentions, which can differ from the page's own. undated means
the age is unknown. Treat it as possibly old.
- "Undated results": most results carried no date although you asked for
recency. Confirm dates on the pages themselves.
- "Gated engines", "Silent engines", "Rate-limited engines", "Benched engines",
rescued via and requested but contributed nothing all mean coverage was
partial. An empty result does not show that nothing exists. Rephrase or
change engines before concluding.
- An engine that "answered with results that match only the first word of the
query" had those results discarded. You need not act on it, but coverage was
narrower.
- The kind shown on a result (
government, academic, news, paper,
forum, code) describes the site. It says nothing about quality.
- An engine error about a missing API key: ignore it and continue with the
keyless engines. Never ask the user for a key.
6. Report
- Give each fact with source and date, e.g. "Registration closes 2026-10-15
(organizer page, updated 2026-09-02, retrieved today)" with the URL.
- Separate what you read on a primary page from what only secondary sources say.
- End with a "Could not verify" list: every requested detail you did not find
on a page you actually read, and every unresolved conflict. Never fill a gap
from a snippet, from memory, or by inference from an earlier edition.
- If nothing authoritative was reachable, say so. Do not fall back silently to
snippet-based answers.
Cost
A normal verification is one search, one or two fetch calls on the primary
page, and one corroborating source. Use research(question, fetch=False) to
choose what to read, keep depth at 2-3, and use cache_search to find a page
you already read this session.
1---2name: verified-research3description: "Use with the free-search MCP tools whenever a web lookup must yield facts someone will rely on: dates, deadlines, prices, prizes, fees, rules, eligibility, schedules, versions, statistics, news, or anything that may have changed. Search results only locate sources; this workflow opens the official or primary page, checks publish date and cache age against today, confirms key facts with a second independent source, and reports dated citations plus what could not be verified. 中文触发:查官网、核实信息、最新消息、报名或截止时间、赛事组别/赛区/奖金、价格、政策、版本号、是否过期或仍有效。"4metadata:5 source: https://skillmd.com/skills/sweetcornna/verified-research6---78# Verified research with free-search910Use search results to find pages, and take facts only from the pages. A snippet11is an engine's summary of a page as it looked when it was crawled: it drops12qualifiers, merges editions, and is often years old. Use this workflow whenever13someone will rely on the answer.1415Tool names below are the bare names on the `search` MCP server; your host may16show them with a prefix.1718## 0. Fix today's date1920Know today's date before judging anything as current. Take it from your system21context. Failing that, the `retrieved …` stamp on a fresh (uncached) `search`22result is the server's clock. "Latest", "this year" and "upcoming" are relative23to it. Do not infer the current year from search results or from memory.2425## 1. Locate: `search` is for URLs only2627- `search(query)` finds candidate pages. Add `freshness="month"` (or28 `"day"`/`"week"`/`"year"`) when recency matters, `include_domains=[...]` once29 you know the official domain, `category=` for papers, filings, news, datasets.30- Some facts have a registry, and the search asks it on its own when the31 question names the thing plainly: a package ("latest fastapi version"), a32 CVE id, two currencies ("100 usd to cny"), a place and a weather word, a33 country and an indicator ("china gdp"), a city and a time word, a domain,34 a coin, an iOS app. The record comes back as the first result with the35 publisher's date, marked "Record sources", and can be cited as it is.36 `category=` ("software", "security", "weather", "finance.fx", "stats",37 "calendar", "reference") asks the same sources explicitly.38- Pick the primary source: the organizer's, vendor's, regulator's or author's39 own page, the filing, the paper, the repository. Aggregators, listicles,40 reposts, forums and AI-written summaries are useful only for finding it.41- Never copy a date, amount, name, rule or number from a snippet or the `Lead:`42 line into your answer.43- Search in the language of the source. A Chinese query is routed to a Chinese44 index automatically; to force more of them use `engines=["so360", "quark", "baidu"]`.4546## 2. Read the source itself4748| Need | Call |49|---|---|50| One page | `fetch(url)` |51| Several candidates | `fetch_batch(urls)` |52| Find and read in one round trip | `research(question, depth=3, include_domains=[...])` |53| Long PDF or DOCX (rules, notices) | `read_doc(source, start, length)`, next page at `start + returned_chars` |54| Dates, prices, event metadata as fields | `extract_structured(url)`: JSON-LD `startDate`, `price`, `datePublished`, `dateModified` |55| One question across 2-5 pages | `compare(question, urls)` |56| A paper you are about to cite | `paper_graph(paper)` for retraction and correction notices |5758Divisions, regions, fee tables, prize breakdowns and deadlines usually sit on a59sub-page or in a linked PDF (rules, FAQ, 章程, 通知). If the landing page lacks60them, follow its links with `fetch` or `read_doc` before you go back to search.61If `fetch` returns a near-empty body, retry with `render="browser"`.6263## 3. Check that it is current6465For every source you rely on, establish three things.66671. When it was published or last updated. Look for `published <date>` in a68 `fetch` header, `datePublished`/`dateModified` from `extract_structured`, or69 a date in the body. `no publication date found` means the page is undated.70 Say so, and do not assume it is recent.712. When this copy was retrieved. Pages and searches are cached for up to 772 days, and `cached N days ago` in the output means you are reading a stored73 copy. For anything that changes (deadlines, prices, availability, standings,74 versions, breaking news) pull it again with `fetch(url, force_refresh=True)`,75 or `max_age_hours=0` on `search` and `research`.763. Which edition it describes. The year, season, version or round on the page77 must match the one asked about. Last year's page for an annual event is the78 most common wrong answer, because it is official and detailed and so looks79 right. If only an older edition is published, report it as that edition and80 say the current one is not announced.8182`freshness=` is best-effort and keeps undated results, so a result that83survives `freshness="month"` may still be old. The search says how many results84it could date (`dated: 3/10`) and marks each one.8586## 4. Corroborate8788- A fact the user will act on needs one primary source, or two independent89 ones. Independent means different origin: two outlets reprinting one press90 release or wire story count as one source. Identical wording usually means a91 shared origin.92- Prefer primary over secondary, and the newer of two primary pages.93- When sources disagree, do not average and do not pick silently. Report both94 values, who says which, their dates, and which you treat as authoritative95 and why.9697## 5. Read the warnings9899- `dated <date>` on a result means the date came from a feed or an API field.100 `<date> (from snippet text)` means it was lifted out of prose and may be a101 date the page mentions, which can differ from the page's own. `undated` means102 the age is unknown. Treat it as possibly old.103- "Undated results": most results carried no date although you asked for104 recency. Confirm dates on the pages themselves.105- "Gated engines", "Silent engines", "Rate-limited engines", "Benched engines",106 `rescued via` and `requested but contributed nothing` all mean coverage was107 partial. An empty result does not show that nothing exists. Rephrase or108 change engines before concluding.109- An engine that "answered with results that match only the first word of the110 query" had those results discarded. You need not act on it, but coverage was111 narrower.112- The kind shown on a result (`government`, `academic`, `news`, `paper`,113 `forum`, `code`) describes the site. It says nothing about quality.114- An engine error about a missing API key: ignore it and continue with the115 keyless engines. Never ask the user for a key.116117## 6. Report118119- Give each fact with source and date, e.g. "Registration closes 2026-10-15120 (organizer page, updated 2026-09-02, retrieved today)" with the URL.121- Separate what you read on a primary page from what only secondary sources say.122- End with a "Could not verify" list: every requested detail you did not find123 on a page you actually read, and every unresolved conflict. Never fill a gap124 from a snippet, from memory, or by inference from an earlier edition.125- If nothing authoritative was reachable, say so. Do not fall back silently to126 snippet-based answers.127128## Cost129130A normal verification is one `search`, one or two `fetch` calls on the primary131page, and one corroborating source. Use `research(question, fetch=False)` to132choose what to read, keep `depth` at 2-3, and use `cache_search` to find a page133you already read this session.