Add x/followers domain skill (followers/profile scraping)

[MoayadAlismail]

Adds agent-workspace/domain-skills/x/followers.md — a playbook for reading other users' followers/following and profile data on X.

Covers:

• Auth check via ct0+twid (auth_token is HttpOnly) and the login-wall redirect.
• The public follower-list pagination cap (observed 73 unique on /followers for a large account, plus ~23 from /verified_followers → ~85–90 unique) — it's a per-request limit, not a soft-block.
• DOM extraction via [data-testid="UserCell"] (preferred over GraphQL), with the small-step (~800px / 1.4s) scroll pattern needed because X virtualizes the list.
• Stable profile-page selectors, GraphQL capture notes, and gotchas.
• Shared-daemon tab-stealing failure mode + the Target.createTarget(url=...) / js(..., target_id=...) pinning fix (hit live when another agent shared the daemon).
• Explicit "don't collect avatars" guidance (no face-based filtering).

🤖 Generated with Claude Code

---

Summary by cubic

Adds agent-workspace/domain-skills/x/followers.md, a practical guide for scraping X followers/following and profile data via the DOM. It clarifies auth checks, the real pagination cap, reliable selectors, and a fix for shared-daemon tab stealing.

• New Features
  • Auth detection via ct0 + twid and the login-wall redirect check.
  • Public list cap documented: ~73 on /followers plus ~20 from /verified_followers (~85–90 unique); following similar.
  • DOM-first approach with stable selectors and small-step scrolling to avoid virtualization misses.
  • GraphQL capture tips (required headers, cursors) and note that replaying does not bypass the cap.
  • Shared-daemon fix using Target.createTarget(...) + js(..., target_id=...); includes guidance to avoid collecting avatars.

^{Written for commit 1281796. Summary will update on new commits.}

[cubic-dev-ai]

1 issue found across 1 file

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="agent-workspace/domain-skills/x/followers.md">

<violation number="1" location="agent-workspace/domain-skills/x/followers.md:67">
P1: Follower-harvesting loop uses `prev` before initialization, causing a `NameError` on the first iteration.</violation>
</file>

<sub>Reply with feedback, questions, or to request a fix.

Fix all with cubic | Re-trigger cubic</sub>

[cubic-dev-ai]

P1: Follower-harvesting loop uses prev before initialization, causing a NameError on the first iteration.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At agent-workspace/domain-skills/x/followers.md, line 67:

<comment>Follower-harvesting loop uses `prev` before initialization, causing a `NameError` on the first iteration.</comment>

<file context>
@@ -0,0 +1,105 @@
+    for c in json.loads(js(EXTRACT)):
+        if c.get("at") and c["at"] not in seen:
+            seen[c["at"]] = c
+    new = len(seen) - prev; prev = len(seen)
+    no_growth = no_growth + 1 if new == 0 else 0
+    js("window.scrollBy(0, 800)")
</file context>