social-scraping-policylisted

# Social Scraping Policy & Operating Manual How to read / scrape X (Twitter) and Xiaohongshu (RedNote) without damaging Bingran's accounts, **and** how to turn a post URL into a card on bingran.ai/posts. Three parts: 1. **Why these platforms are different** — the threat model, so the rules below make sense. 2. **The operating manual** — concrete, numeric defaults: account, tool, pacing, fingerprint, signals, backoff. 3. **The /posts pipeline** — concrete recipes that take a post URL and land a card on bingran.ai (data layout, per-platform extractors, thumbnail strategy, "paste-a-link" flow). This skill is the gate AND the playbook. If you read it and still don't know what to do, stop and ask Bingran. ## Canonical index — code & docs that belong to this skill Everything you'd reach for when handling these tasks. The skill is the single source of truth; AGENTS.md just points here. | Asset | Where it lives | Why it isn't inside this skill dir | |---|---|---| | `add-post.mjs` — URL → metadata → JSON-append script | `personal-site/scripts/add-post.mjs` | Wired into `personal-site/package.json` as `npm run post:add`; moving it would break that workflow. Treat *this* SKILL.md as the docs of record; `--help` in the script is intentionally terse. | | `posts.json` — the data | `personal-site/content/posts/posts.json` | It's the personal-site's content, not a skill artifact. | | `post-card.tsx` — render component | `personal-site/components/post-card.tsx` | Same — site code. | | T