agentleFS
glideaglidea/supersearch-skills/skills/super-search/SKILL.md

super-search

Skill4 starsChanged 3 months agoMITView on GitHub

Aggregate Web, Linux.do, V2EX, Xiaohongshu, WeChat Official Accounts, Douyin, Telegram, Reddit, and X. Search every source by default or only user-selected sources for research, fact-checking, product comparison, technology choices, troubleshooting, and trend tracking with primary-source links.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. Hudson7777/claude-os★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository9 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use Chinese queries for Chinese communities and add natural English queries for Reddit and X.

    line 13
  2. Very specific:

    When the user does not specify sources, run $linuxdo-search, $v2ex-search, $xiaohongshu-search, $wechat-search, $douyin-search, …

    line 14
  3. Very specific:

    When Linux.do, V2EX, or Reddit returns no citation, immediately run built-in Web searches for site:linux.do "core name", site:v2ex.com/t …

    line 16
  4. Very specific:

    When X returns citations, cite only those links.

    line 27
  5. Very specific:

    When any community result returns search_executed: false, mark that source as not successfully searched and do not use generated links from …

    line 28
  6. Distinctive:

    Place links next to important claims.

    line 29

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Super Search

Search workflow

  1. Load and execute the dependent source skills before searching. This skill is an orchestrator, not a replacement for $linuxdo-search, $v2ex-search, $xiaohongshu-search, $wechat-search, $douyin-search, $telegram-search, $reddit-search, and $x-search. When a source skill is listed below, read its SKILL.md and invoke that skill's search workflow or tool. Do not treat merely reading this file as having searched the source. Do not silently replace a dedicated source search with built-in Web search.
  2. Identify the core question, time range, and claims that need verification.
  3. Use Chinese queries for Chinese communities and add natural English queries for Reddit and X. Do not merely translate word for word.
  4. When the user does not specify sources, run $linuxdo-search, $v2ex-search, $xiaohongshu-search, $wechat-search, $douyin-search, $telegram-search, $reddit-search, $x-search, and built-in Web search in parallel without skipping sources. When the user specifies sources, call only those sources. Before marking any dedicated source unavailable, retry the actual dedicated skill/tool three times, including HTTP 4xx/5xx responses and connection errors. If the skill or tool cannot be loaded or invoked, report that exact failure and do not claim the source was searched. Only then use the documented Web fallback for Linux.do, V2EX, or Reddit. When Linux.do, V2EX, or Reddit returns no citation, immediately run built-in Web searches for site:linux.do "core name", site:v2ex.com/t "core name", or site:reddit.com "core name". If dedicated and Web results conflict, prefer an accessible primary post.
  5. For Web search, prefer official documentation, project repositories, papers, announcements, and primary statements. Use community search to find practical experience, disputes, failures, and user feedback.
  6. For real-time topics, pass the user's --from and --to values to X. When no range is given, choose a reasonable range and state it.
  7. Merge duplicate events and reposts. Trace claims to the earliest or most authoritative primary source instead of citing aggregators.
  8. Compare sources. Distinguish confirmed facts, community opinions, and unverified information. Do not treat popularity as evidence.
  9. By default, show local cover images for the most relevant Douyin and Xiaohongshu results, with no more than three images total. Link each image to its source post. When the user requests no covers or text-only results, tell those sources to skip covers and omit covers from the final answer.

Citation rules

  • Cite only pages actually returned by the Web tool.
  • Prefer citations returned by Linux.do, V2EX, and Reddit. When built-in Web fallback runs, cite platform posts actually returned by Web. For Xiaohongshu, WeChat, and Douyin, cite only each result's url; when a WeChat original cannot be opened, use sogou_url from the same result. For Telegram, prefer original message links returned by get_message_link; otherwise provide the chat name, timestamp, and message ID.
  • When X returns citations, cite only those links. When search_executed is true without citations, summarize content and state that the upstream service provided no citable source posts. Do not describe it as no discussion.
  • When any community result returns search_executed: false, mark that source as not successfully searched and do not use generated links from it.
  • Place links next to important claims. State when no reliable evidence was found.

Output format

Answer directly first, then include as needed:

  • Confirmed facts: conclusions supported by official or multiple independent sources.
  • Community observations: representative experience and disagreements from Linux.do, V2EX, Xiaohongshu, WeChat, Douyin, Telegram, Reddit, and X.
  • Uncertainty: conflicts, sampling bias, stale information, and search failures.

Stay concise. Do not dump results source by source or include irrelevant material merely to cover every source.

End every response with a ## References section. Keep links next to supported claims and repeat every used link in this final section. Use one line per website, preserve the result title, deduplicate links, and separate multiple links with commas. Use the website or platform name as the label, not Web:

## References

Linux DO: [result title](https://example.com), [result title](https://example.com)
OpenAI: [result title](https://example.com)

Include only websites with references actually used in the answer. For Telegram private messages without shareable links, use Telegram: chat name / timestamp / message ID. When no citable source exists, write No citable sources. below the heading.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.