Labali Weibo Download User Posts

labali-weibo-download-user-posts

Treat this skill as a layered system, not a single script.

Layer Contract

1. SKILL.md (this file) is the policy layer.
   • Define goals, constraints, success criteria, and decision boundaries.
   • Stay semantic and stable across UI changes.
2. references/architecture.md is the strategy layer.
   • Define execution model, failure handling, and quality standards.
3. scripts/*.ts is the execution layer.
   • Scripts are execution assets, not the skill definition itself.
   • Use persistent browser profile for manual-login reuse.
   • Keep extraction and downloading logic modular and replaceable.

Required Constraints

• Use browser automation only.
• Do not use private/undocumented Weibo APIs.
• Reuse manual-login session via unified Chrome CDP startup: open -na "Google Chrome" --args --remote-debugging-port=9223 --user-data-dir="$HOME/.chrome-labali-no-proxy" --no-proxy-server.
• Prefer semantic extraction from visible page state and loaded resources.
• Download target user's timeline posts and assets (text, images, optional videos).
• Export metadata files: posts.json, user.md, and per-post post.md.
• Keep partial success: keep downloaded files even if some URLs fail.

NEVER

• Never delete or discard partially downloaded files on URL failure — keep all successfully downloaded content.
• Never report success unless at least one post was extracted and the output folder was created with posts.json and user.md.
• Never report success based on action completion alone — verify the output folder and metadata files exist.

Success Criteria

A run is successful only when all conditions hold:

1. A user output folder is created under target local directory.
2. Folder naming format is <timestamp>-<user_slug>.
3. posts.json and user.md are generated.
4. At least one post is extracted from timeline.
5. Per-post folders contain post.md and downloaded media when available.
6. URL output and logs retain canonical user URL form.

Runtime Inputs

Use skill.yaml as the source of truth for input schema.

Operational Mode

• Default mode: guided browser flow + semantic extraction + authenticated media download.
• Startup guidance:
  • launch/reuse Chrome by the unified open -na "Google Chrome" CDP command,
  • connect via CDP port,
  • open target user page,
  • check login status,
  • if required, guide user to complete manual login in the same window.
• Input guidance:
  • if user_url is missing, prompt interactively,
  • if output_dir is missing, prompt interactively with default.
• Pagination guidance:
  • scroll timeline,
  • click visible expansion controls,
  • stop when no more content appears or explicit end markers are present.

If failure handling or extraction decisions are unclear, load references/architecture.md. If extraction strategy for a specific content type is unclear, load references/plan.md.

Resources

• Architecture and standards: references/architecture.md
• Workflow map and extraction plan: references/plan.md
• Shared runtime and downloader helpers: scripts/core.ts
• Main orchestration: scripts/executor.ts
• CLI entry: scripts/run.ts
• Regression checks: tests/test_regression.sh