xiaohongshu-account-research
About xiaohongshu-account-research
A research skill for comparing Xiaohongshu accounts using evidence from recent posts. It aggregates post data into account snapshots, rankings, top-note links, and recurring title patterns, while flagging unavailable profile fields.
Key features
- 01Samples recent posts from profile URLs
- 02Aggregates account-level post evidence
- 03Ranks accounts by performance and thematic fit
- 04Surfaces top notes and repeated title patterns
- 05Flags unavailable profile data
Use cases
- 01Shortlisting benchmark accounts
- 02Comparing competitor or partner accounts
- 03Finding top-performing notes and title patterns
- 04Reviewing account content mix from recent posts
Install
Copy the command and run it in a terminal.
npx skills add https:/ / github.com/ postplusai/ postplus-skills --skill xiaohongshu-account-researchAfter installing, type/xiaohongshu-account-researchin Claude Code to confirm it loaded.
Bundled files
References and scripts installed along with SKILL.md.
_postplus_shared/00-core/shared-collection/scripts/collection_actor_run.mjs_postplus_shared/00-core/shared-collection/scripts/lib/hosted_collection_bridge.mjs_postplus_shared/00-core/shared-runtime/scripts/download_videos_from_manifest_with_ytdlp.mjs_postplus_shared/00-core/shared-runtime/scripts/lib/execution_approval.mjshosted_billing_summary.mjshosted_execution_protocol.mjshosted_media_generation_bridge.mjslocal_skill_cli.mjsnetwork_runtime.mjspostplus_cli_config.mjspostplus_cloud_client.mjsskill_runtime.mjs_postplus_shared/scripts/lib/campaign-report-manifest.mjsreferences/actor-selection.mdnormalized-schema.mdscripts/analyze_xhs_accounts.mjsbuild_xhs_account_actor_input.mjsnormalize_xhs_account_dataset.mjsrank_xhs_accounts.mjsscripts/lib/account_common.mjsDetails
From the project READMEXHS Account Research
Follow shared public skill rules in:
postplus-sharedpublic skill rules
Legacy alias: xhs-account-research.
Use this skill when the user wants to:
- compare one or more Xiaohongshu accounts
- build a shortlist of benchmarkable accounts from known profile URLs
- understand what an account usually posts, what title patterns repeat, and which notes perform best
- produce account-level snapshots for partner or competitor review
Read these references before implementation:
skills/20-research/xhs-account-research/references/actor-selection.mdskills/20-research/xhs-account-research/references/normalized-schema.md
Default posture
Start from validated recent-post evidence.
Default actor:
rednote-xiaohongshu-user-posts-scraper
Do not default to rednote-xiaohongshu-profile-scraper.
Treat it as non-default until it returns usable profileData on the public skill surface.
Before collection, tell the user that profile bio, follower count, and other profile-level fields are not reliable on the released path. Results are based on recent posts unless the validated dataset exposes them.
What this skill is for
- recent-post sampling from known profile URLs
- account-level aggregation from post evidence
- ranking accounts by observed post performance and thematic fit
- producing account snapshots with top notes and repeated title patterns
What this skill is not for
- keyword discovery
- comment mining
- follower counts
- bio extraction
- posting cadence claims when no publish timestamps are available
Failure posture
- fail if no
profileUrlsorprofileIdsare provided - fail if the validated actor returns zero posts
- fail if account aggregation cannot produce any top note URL
- report missing profile-level fields explicitly instead of inventing them
Public Skill Execution Contract
- keep account briefs, actor inputs, raw datasets, normalized datasets,
rankings, and analysis outputs under
<work-folder>/.postplus/xiaohongshu-account-research/ - keep only final user-facing account shortlists or reports outside
.postplus/ - start with a bounded first pass, usually
1-3accounts and8-15recent posts per account - if PostPlus Cloud service is unavailable, unauthorized, or returns a stable network error, stop immediately instead of switching to ad hoc shell glue
Main scripts
scripts/build_xhs_account_actor_input.mjsscripts/normalize_xhs_account_dataset.mjsscripts/rank_xhs_accounts.mjsscripts/analyze_xhs_accounts.mjs
Use the shared collection runner for actor calls:
${CLAUDE_SKILL_DIR}/_postplus_shared/00-core/shared-collection/scripts/collection_actor_run.mjs
Minimal workflow
node ${CLAUDE_SKILL_DIR}/scripts/build_xhs_account_actor_input.mjs \
--brief <work-folder>/.postplus/xhs-account-brief.json \
--output <work-folder>/.postplus/xhs-account-actor-input.json
node ${CLAUDE_SKILL_DIR}/_postplus_shared/00-core/shared-collection/scripts/collection_actor_run.mjs \
--collection-key xhs-account-posts \
--input <work-folder>/.postplus/xhs-account-actor-input.json \
--output <work-folder>/.postplus/xhs-account-raw.json
node ${CLAUDE_SKILL_DIR}/scripts/normalize_xhs_account_dataset.mjs \
--input <work-folder>/.postplus/xhs-account-raw.json \
--output <work-folder>/.postplus/xhs-account-normalized.json
node ${CLAUDE_SKILL_DIR}/scripts/rank_xhs_accounts.mjs \
--input <work-folder>/.postplus/xhs-account-normalized.json \
--theme "workplace,office workers" \
--output <work-folder>/.postplus/xhs-account-ranking.json
node ${CLAUDE_SKILL_DIR}/scripts/analyze_xhs_accounts.mjs \
--input <work-folder>/.postplus/xhs-account-normalized.json \
--output <work-folder>/.postplus/xhs-account-analysis.json
Good output
Return:
- account shortlist
- top note URLs per account
- median and max like signal
- content-type mix
- repeated title-pattern families
- cover-shape mix
- explicit data-quality warnings for unavailable profile-level fields