juejin-publish-and-data

name	juejin-publish-and-data
description	Saves Markdown articles to Juejin drafts and fetches creator analytics using Playwright storageState.json or cookie JSON. Supports recent content analytics, 7-day article trends, and follower analytics.
metadata	{"openclaw":{"requires":{"anyBins":["node","npx"]}}}

编排入口：多 skill 组合流程请使用 media-manager 总控 skill 与 media CLI。本 skill 为 deep-dive。

Juejin Publish And Data

Language

Match the user's language.

CLI（推荐）

执行前先 media workspace show。Auth：$WORKSPACE/.media-manager/auth/juejin/。

操作	命令
保存草稿	media juejin post --file output/{slug}/article.md --draft
抓取运营数据	media juejin analytics fetch
导出登录态	media juejin auth export
检查登录	media juejin auth check

Script Directory（Deep-dive）

Determine this SKILL.md directory as {baseDir}. Prefer media juejin ... above.

What This Skill Does

This skill supports two separate capability lines for Juejin creator workflows:

• save Markdown articles into Juejin editor drafts
• fetch creator analytics from content data and follower data pages
• verify whether the current auth state can enter creator pages without relogin

Output layers for fetch:

• raw capture layer: original network payloads and fallback page state for debugging
• normalized JSON layer: compact LLM-friendly output for content overview, latest 5 articles, article links, recent 7-day article trends, and follower metrics

Fetch result location:

• if --output is provided, write results into that directory
• else use default_output_dir from EXTEND.md when configured
• else use ./juejin-data-output
• relative output paths are resolved from the shell working directory where the command is invoked
• always treat the final JSON file path printed/reported by the script as the source of truth for downstream LLM analysis

This skill does not perform login automation, QR code handling, SMS verification, CAPTCHA bypass, scheduling, or multi-account orchestration.

Required Inputs

• Preferred auth file: Playwright storageState.json captured from an already logged-in Juejin browser session
• Fallback auth file: cookie JSON exported from the same logged-in browser session
• For publishing: a Markdown file, optionally with frontmatter title/tags/cover/column/visibility
• Optional CLI overrides: --title, --cover, --tags, --column, --draft

Auth file discovery order:

1. CLI --state / --cookie
2. $WORKSPACE/.media-manager/auth/juejin/storageState.json
3. $WORKSPACE/.media-manager/auth/juejin/cookies.json
4. <skillRoot>/.auth/（兼容）

Preferences (EXTEND.md)

Optional config file path order:

1. /.config/EXTEND.md
2. /EXTEND.md

Supported keys:

• default_page: content | follower | both
• default_output_dir: output folder path
• default_post_mode: draft
• default_tags: comma-separated tags
• default_column: default column name
• default_visibility: default visibility label text
• default_save_raw: true | false
• default_timeout_ms: timeout in milliseconds
• cookie_file_name: cookie file name
• storage_state_file_name: storage state file name

First-time setup guide: references/config/first-time-setup.md

Storage state export guide: references/cookie/export-storage-state.md

Output format guide: references/output-format.md

Article posting guide: references/article-posting.md

Workflow

Execution policy:

• first run on a machine, fresh environment, or after dependency/browser changes: run Step 1, Step 2, then Step 3
• recurring runs with the same installed dependencies, same browser setup, and still-valid auth state: skip directly to Step 4A or Step 4B
• rerun Step 3 whenever storageState.json or cookies.json changed, the session may have expired, or creator access is uncertain
• rerun Step 1 and Step 2 only when environment validation is needed again

First-run or recovery checklist:

Progress:
- [ ] Step 1: Install dependencies and browser
- [ ] Step 2: Run environment check
- [ ] Step 3: Verify creator login state
- [ ] Step 4A: Fetch creator analytics if needed
- [ ] Step 4B: Post article if needed
- [ ] Step 5: Report output files or draft result

Step 1: Install dependencies and browser

Run in {baseDir}/scripts:

npm install
npx playwright install chromium

For Ubuntu missing libraries, see references/ubuntu/headless-setup.md.

Step 2: Environment check

npx tsx {baseDir}/scripts/check-environment.ts

Step 3: Login state check

media juejin auth check

If redirected to login page or creator page body shows relogin content, stop and refresh login state.

Step 4A: Fetch analytics

Recommended command:

media juejin analytics fetch

Supported options:

• --page content|follower|both
• --start YYYY-MM-DD
• --end YYYY-MM-DD
• --output
• --state
• --cookie
• --save-raw / --no-save-raw (default: off)
• --probe
• --headful
• --timeout

Step 4B: Post article

Default draft-first command:

media juejin post --file output/{slug}/article.md --draft

Supported options:

• --file <article.md>
• --title
• --cover
• --tags <tag1,tag2>
• --column
• --visibility
• --draft
• --state
• --cookie
• --headful
• --timeout

Step 5: Report results

For fetch, always report:

• absolute output JSON file path
• whether raw data folder is generated
• normalized latest article count and follower rows count
• whether content overview, article links, and 7-day trends were captured
• tell the user to read the reported JSON path as the final fetch result source

For post, always report:

• final mode: draft
• title used
• article URL if available
• article ID if available
• tags/column/cover actually applied
• any optional fields that could not be confirmed through UI

Safety And Boundaries

• Do not ask for account password, SMS code, or QR login secrets.
• Do not implement login bypass or CAPTCHA bypass.
• Treat cookie JSON and storageState.json as sensitive files and never commit them.
• Only save drafts. Do not attempt final publish from this skill.