x-account-pitfalls

Account Health and Throttling Recovery

R — Original Text (Reading)

"Frequently switching IPs or devices easily triggers verification or a temporary label, usually lifted within 24h. Mass following, unfollowing, and replying looks like a script, and in severe cases leads to throttling and account bans. Changing profile info during Blue V review triggers re-review. Check throttling: use another account to search from:@username; if it can't be found = throttled. First check Privacy settings; if that doesn't work, write an appeal email to @Support and do no account operations for 48h. For a temporary label, don't touch the account first; it ends in a few hours."

— Mumaren, X Account Launch Pitfall Avoidance Guide · 7 Major Pitfalls; AIP Going Global Tutorial · 5.1.2 Account Check

I — Methodology Skeleton (Interpretation)

Maintaining X account health has two layers:Prevention(don't let risk control trigger) andRecovery(how to save it after triggering).

Prevention layer — avoid the 7 major pitfalls:

Use a fixed IP and device, don't switch frequently (triggers verification/temporary label)
Don't mass follow/unfollow/reply (operations look like a script = judged as a bot)
Don't change avatar/nickname/bio during Blue V review (changing profile info triggers re-review, Blue V is lost)
Watch the data and actively interact for the first hour after posting a tweet (don't just post and ignore it)
Don't force-hijack traffic (posting irrelevant content under high-traffic posts = getting reported)
Build content capability, don't only chase traffic tricks (platform tricks change, content capability is transferable)
Don't fear a ban, appeal more (appealing is a regular recovery method, not a dead end)

Recovery layer — throttling self-check and appeal:

Self-check: use another account to search from:@你的用户名, if it can't be found = throttled
Distinguish two levels: temporary label (applied when operations look like a script, automatically lifted in a few hours~24h, during which don't touch the account) vs formal label/throttling (posts can't be found, requires writing an appeal email + no account operations for 48h)
After appealing, review the trigger cause (switching IP/mass operations/changing profile info during review/unbalanced following-to-follower ratio), and prevent recurrence accordingly

Core insight: throttling does not equal a ban — a temporary label can self-heal (wait 24h without touching it), and a formal label can be appealed (write an email + 48h of silence). Continuing to operate in a panic instead worsens risk control.

A1 — Application in the Book (Past Application)

Case 1: Mumaren's 11,897 followers wiped out overnight, recovered after 12 appeals

Problem: The account was suddenly banned, 11,897 followers wiped out overnight, all content disappeared
Use of the methodology: Pitfall 7 of the 7 major pitfalls, "Don't fear a ban, just appeal more" — don't give up, repeatedly submit appeals to @Support. With appropriate intervals between each appeal, persist for 12 rounds
Conclusion: A ban is not the end; appealing is a regular recovery method; but it requires patience, and a single appeal may not succeed
Result: After 12 appeals, the account was finally recovered

Case 2: Mumaren restarted from 0 to 4,200 in 21 days, launch post got 280,000 impressions

Problem: After the old account was recovered, its data was already ruined, and it needed to be relaunched from 0
Application of the methodology: After recovery, don't rush to go back to the old playbook; instead, use "watch the data for 1 hour after posting + proactively ask people to help repost" to get through the cold start. The launch post relied on friends helping to repost, one piece of content got 280,000 impressions, and followers grew from 16 to over 700
Conclusion: When restarting after being throttled/banned, the core is relying on friends to help repost to get through the golden window of the cold start, rather than relying on bulk operations to gain followers quickly (that will trigger risk control again)
Results: Went from 0 to 4,200 followers in 21 days, launch post got 280,000 impressions

A2 — Trigger scenarios (Future Trigger) ★

In what situations will users need this skill?

Tweets suddenly can't be found / impressions drop off a cliff — suspect being throttled, need self-check and troubleshooting
Account is notified of unusual activity / temporary label — don't know what to do, or whether it will recover on its own
Worried that a certain action (switching IP node / bulk following / changing profile during review period) will trigger risk control — need to know where the red lines are
Account has been banned/frozen and needs to appeal for recovery — need the appeal process and copy templates
Have been throttled/banned before and want to prevent recurrence — need to establish a daily operation safety checklist

Language signals (activate when these appear in the user's words)

"Tweets can't be found" / "being throttled" / "shadowbanned" / "tweets not showing up"
"account abnormal" / "triggered risk control" / "temporary label" / "account flagged"
"judged as a bot" / "account suspended" / "what to do if banned"
"how to appeal" / "appeal suspension" / "how to write an appeal email"
"what happens if I switch IP frequently" / "is bulk following safe" / "can I change profile during review period"
"follow to follower ratio" / "follow to follower ratio"

Distinction from adjacent skills

Difference from x-account-setup : setup is "prevention" — pre-launch preparation before posting content (profile setup + blue check + account nurturing); this skill is "treatment" — troubleshooting and appeals after problems have already occurred (throttling/ban/label). If the user says "how to prepare on the first day of a new account" → setup; if the user says "what to do if I'm throttled" → this skill.
Difference from x-cold-start-playbook : cold-start is "customer acquisition" — how to gain traction without followers (six strategies + funnel diagnosis); this skill is "life-saving" — recovery when account health has problems. If the user says "posted a tweet but got no traffic" → cold-start; if the user says "tweets can't be found" → this skill.
Difference from x-comment-engagement : engagement is about comment interaction strategies (how to reply with high quality); this skill is about operational red lines (bulk replying = script characteristics, triggers risk control). If the user says "how to write good comments" → engagement; if the user says "replied too much and got judged as a script" → this skill.

E — Executable steps (Execution)

When the skill is activated, the agent should execute the following steps:

Throttling self-check
- Use another account (a friend's account or a alt account; searching with your own account may be inaccurate) to search in the search box for from:@你的用户名
- Can't find it = being throttled; can find it = not being throttled
- If it can't be found, first check whether Settings → Privacy and Safety → Your Posts is turned off
- After confirming the settings are normal, post a new tweet, then use another account to search for from:@你的用户名, and see whether it can be found
- Completion standard: confirm whether the account is being throttled
Distinguish label types and handle them
- Temporary label(Hint: there is a label but the tweet can still be found): Do not touch the account; it will be automatically lifted in a few hours to ~24h. During this period, do not like/comment/follow/post tweets
- Formal label/throttling(Tweet cannot be found): Proceed to step 3 appeal process
- Stop condition: If the self-check result is no throttling (tweets can be found), skip to step 4 to investigate potential risks
- Completion criteria: Clearly determine whether it is a temporary label or formal throttling, and take the corresponding action

Appeal process (throttling/account ban)

Find @Support to appeal, using the following copy template (replace the username):

Dear X Support Team,
I am writing to appeal the suspension of my personal X account, @【YourName】. I was recently informed that my account has been suspended, but I believe this may have been an error. As a personal user, I have always adhered to X's community guidelines and rules. Upon reviewing the suspension notification, I could not find any evidence of violating these policies. I kindly request that you investigate this matter further and reinstate access to my account if possible. Thank you for your time and consideration in resolving this issue.
Sincerely,
@yourname

After submitting the appeal, do not perform any account operations within 48h (do not post tweets/like/comment/follow/unfollow)
A single appeal may not succeed; you can submit multiple times (in the Trojan Horse case, it took 12 appeals to recover the account)
Completion criteria: The appeal has been submitted, and the 48h silent period has begun

Investigate the trigger cause
- Frequently switching IP/device? → Fix the IP and device, and stop frequently switching nodes
- Bulk following/unfollowing/replying? → Stop all bulk operations and switch to manual, one-by-one actions
- Changed profile during blue check review? → Wait for the review to end, and absolutely do not touch the profile during the review period
- Following count far exceeds follower count? → Stop following and let the ratio naturally return to balance
- Excessive follow-for-follow (more than 15 per day)? → Control the follow-for-follow frequency, no more than 15 people per day
- Completion criteria: Identify at least 1 possible trigger cause
Prevent recurrence — Daily operation safety checklist
- Fix the IP and device, and do not switch frequently
- Do not bulk follow/unfollow/reply; keep random intervals for any operation (simulate a real person's rhythm)
- No more than 15 follow-for-follow per day
- The following count should not far exceed the follower count (e.g., following 100 people but having only 1 follower = dangerous)
- Do not change the avatar/nickname/bio during the blue check review period
- Watch the data and actively interact in the 1 hour before posting a tweet (do not just post and ignore it)
- Do not force yourself into traffic (only post high-quality relevant comments under high-traffic posts)
- Platform policy red lines/sensitive topics must be avoided; for the specific list, refer to X's official community guidelines
- Build content capability; do not treat exploiting loopholes as a long-term strategy (platform tricks will change, but content capability is transferable)
- Completion criteria: The user has established awareness of daily operation safety and knows which operations will trigger risk control

B — Boundary ★

Do not use this skill in the following situations

Posting content normally and the account has no problems — this skill is "treatment + risk control prevention," not daily content operations. You should switch to a content creation skill
Preparing for launch after registering a new account/unfreezing an old account (profile setup + blue check + account nurturing) — you should switch to x-account-setup(That is the "prevention" launch checklist; this skill is the "treatment" after problems occur)
Posted content but has no traffic/no followers — you should switch to x-cold-start-playbook(That is a "customer acquisition" problem, not an "account health" problem. Key test: if the tweet can be found with from:@ search = not throttled = not a problem for this skill)

Failure modes the author warns about in the book

ce02 — Frequent IP switching triggers risk control: Frequently switching VPN nodes or login devices triggers platform risk control; the account gets temporarily rate-limited or tagged, and in severe cases banned. X judges user authenticity through IP stability and device fingerprinting. Prevention: use a fixed IP and device.
ce03 — Batch operations flagged as scripts: Batch following/unfollowing/replying, with highly regular operation patterns (uniform intervals, copy-pasted content), is judged by the anti-cheat system as scripted behavior, leading to rate-limiting/banning. Prevention: maintain the randomness of manual operations; don't do things in batches.
ce17 — Imbalanced following-to-followers ratio flagged as a follow-for-follow bot: When the number of accounts followed far exceeds the number of followers (e.g., following 100 people but having only 1 follower), the system judges it as a follow-for-follow bot and the account gets rate-limited. Prevention: don't let the number of accounts followed far exceed the number of followers.
ce21 — Game-theory tactics fail overall once the platform tightens up: Follow-for-follow + grabbing comment sections + matrixing are essentially game-theory tactics; once the platform tightens up (cracking down on follow-for-follow, downranking quotes, restricting homogeneous reposting), the whole playbook may fail. Don't treat exploiting loopholes as a long-term strategy — what's truly transferable is content ability itself.
ce07 — Only chasing traffic without building content ability: Being obsessed with traffic-chasing tricks and platform rules while neglecting content ability building. Once the algorithm changes, all the tricks fail and the account goes back to zero. Platform tricks will change; content ability is the thing you can take with you.

The author's blind spots / era limitations

The temporary tag "lifted within 24h" is an empirical value, not an official public commitment from X; the actual recovery time may be longer
Appeal success rates vary by case — Trojan Horse Man succeeding on the 12th appeal is survivor narrative; cases with equal effort that didn't get restored are not shown
The rate-limit self-check method (searching from:@username) may change as the platform's search function is adjusted
"No more than 15 follow-for-follows per day" is an empirical threshold, not an official public rule; the platform's anti-cheat threshold may be adjusted at any time
Game-theory tactics (controlling the following-to-followers ratio, controlling follow-for-follow frequency) are essentially exploiting the gaps in the current anti-cheat system; platform anti-cheat keeps evolving, and what works today may fail tomorrow (ce21)
Platform policy red lines/sensitive topics must be avoided; the specific list is subject to X's official community guidelines, and this skill does not enumerate specific sensitive areas

Adjacent methodologies that are easy to confuse

Rate-limited vs no traffic: Rate-limiting means the account has been downranked by risk control (tweets can't be found via from:@ search); no traffic may be due to insufficient content quality/cold start (findable via from:@ but with low engagement). The former uses this skill to troubleshoot and appeal; the latter uses x-cold-start-playbook to diagnose the funnel.
Temporary tag vs formal tag: A temporary tag is a short-term mark applied when operations look like a script (self-heals in a few hours to 24h; just don't touch the account); a formal tag/rate-limit is more serious (posts can't be found, requires appeal + 48h silence). Conflating the two leads to wrong actions — continuing to interact during a temporary tag worsens risk control, while a formal tag cannot be recovered without an appeal.

Related skills

depends-on x-account-setup: Rate-limit recovery depends on getting launch preparation right — if prevention is done right (fixed IP/no batch operations/no touching profile info during the review period), risk-control risk is greatly reduced, and if problems arise it's easier to pinpoint the trigger cause.
composes-with x-account-setup: Launch preparation (prevention) + rate-limit recovery (treatment) used together, integrating disease prevention and treatment.

Audit information

Verification passed: V1 ✓ / V2 ✓ / V3 ✓
Test pass rate: To be tested (see test-prompts.json / test-results.md for details)
Distillation time: 2026-07-14