agentleFS
gaofeng21cngaofeng21cn/opl-skills/skills/xiaohongshu-repo-scout/SKILL.md

xiaohongshu-repo-scout

Skill0 starsChanged 3 days agoApache-2.0View on GitHub

Use when turning XiaoHongShu or Chinese social-media posts into a verified GitHub repository shortlist, especially for academic, research, AI, data, visualization, writing, or automation tools. Use when the user asks to find repos from XiaoHongShu posts, compare social recommendations against GitHub evidence, check repo stars/freshness from post leads, or confirm whether an OpenCLI/Agent Reach XiaoHongShu lead is actionable. Not for generic web research, posting/publishing, social engagement growth, or repo evaluation without a social-source lead.

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 9 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill to convert social discovery into a defensible GitHub shortlist.

    line 10
  2. Distinctive:

    Check the active XiaoHongShu backend before reading:

    line 16
  3. Distinctive:

    Use the backend established by the current readback.

    line 23
  4. Distinctive:

    Search XiaoHongShu with a query that names the domain and repository/tool intent:

    line 26
  5. Very specific:

    Read promising notes by passing the full search-result URL that includes xsec_token, not a bare note id:

    line 32
  6. Very specific:

    If search works but note reads fail, first check whether the URL lost xsec_token.

    line 38
  7. Distinctive:

    Verify each candidate with GitHub:

    line 46
  8. Distinctive:

    Use current GitHub evidence for stars, archive state, description, and update recency.

    line 53
  9. Distinctive:

    Do not publish, delete, follow, unfollow, like, or comment on XiaoHongShu unless the user explicitly asks and confirms the exact write …

    line 89
  10. Distinctive:

    Treat OpenCLI XiaoHongShu write commands as browser UI automation, not stable official APIs.

    line 90
  11. Distinctive:

    Prefer drafting copy or preparing a confirmation prompt over unattended posting.

    line 91
  12. Very specific:

    Do not claim Agent Reach or OpenCLI is fully healthy from doctor alone.

    line 92

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

XiaoHongShu Repo Scout

Overview

Use this skill to convert social discovery into a defensible GitHub shortlist. Social engagement is only a lead source; a repository becomes a candidate only after exact identity extraction and fresh GitHub verification.

This skill complements agent-reach: use agent-reach or OpenCLI to fetch XiaoHongShu content, then apply this workflow to avoid guessing repo identities or treating post popularity as quality evidence.

Evidence Order

  1. Check the active XiaoHongShu backend before reading:
agent-reach doctor --json
opencli daemon status

Use the backend established by the current readback; public workflow source does not carry one machine's previous backend choice.

  1. Search XiaoHongShu with a query that names the domain and repository/tool intent:
opencli xiaohongshu search "<query>" -f yaml
  1. Read promising notes by passing the full search-result URL that includes xsec_token, not a bare note id:
opencli xiaohongshu note "<full-search-result-url-with-xsec-token>" -f yaml

If search works but note reads fail, first check whether the URL lost xsec_token.

  1. Extract only explicit repository identities:
    • Direct GitHub URLs.
    • Owner/repo strings.
    • Unique project names that can be matched by fresh GitHub search.
    • Exclude posts that only imply a tool, show screenshots, or use vague names without a stable repo identity.
  2. Verify each candidate with GitHub:
gh repo view <owner>/<repo> --json nameWithOwner,description,stargazerCount,forkCount,updatedAt,url,isArchived,defaultBranchRef
gh search repos "<project name>" --sort stars --limit 10

Use current GitHub evidence for stars, archive state, description, and update recency. If GitHub API rate limiting blocks verification, report that as a verification gap and do not claim current stars or freshness.

Candidate Rules

Recommend a repository only when all are true:

  • A XiaoHongShu/source note or search result produced the lead.
  • The exact GitHub repo identity is confirmed.
  • Fresh GitHub metadata was read in the current task.
  • The repo's purpose matches the user's requested domain.
  • Any safety or write-operation boundary is clear.

Skip or downgrade a repository when:

  • The post has engagement but no exact repo identity.
  • Multiple repos could match and none is clearly the one referenced.
  • The repo is archived, stale, unrelated, low-signal, or mostly empty.
  • Verification depends only on a social post, screenshot, README claim, or cached summary.

Output Shape

Lead with the shortlist, then show why each item survived filtering.

For each recommended repo include:

  • Repository name and URL.
  • Why it is relevant to the user's requested workflow.
  • Fresh stars and update signal from GitHub.
  • Source lead: XiaoHongShu query/note title or URL when available.
  • Confidence: high / medium / low.
  • Caveat or next check if the repo should be tried before adoption.

Also include a short rejected/uncertain list when useful, especially for high-engagement posts that failed repo identity or GitHub verification.

Safety Boundaries

  • Do not publish, delete, follow, unfollow, like, or comment on XiaoHongShu unless the user explicitly asks and confirms the exact write action.
  • Treat OpenCLI XiaoHongShu write commands as browser UI automation, not stable official APIs.
  • Prefer drafting copy or preparing a confirmation prompt over unattended posting.
  • Do not claim Agent Reach or OpenCLI is fully healthy from doctor alone; prove the actual channel with a real search/note read.

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.