Safety systems: warmup, pacing, and the ban-risk score
How SkyFollowing protects accounts: the 14-day warmup ramp, human pacing, the 0-100 ban-risk score with every factor documented, and automatic holds.
Safety in SkyFollowing is layered, always on, and mostly invisible: a warmup ramp for young accounts, human-shaped pacing for every action, a continuously recomputed ban-risk score, and an automatic hold when that score turns critical. This page documents each layer with its real numbers.
The warmup ramp
Newly connected accounts run a 14-day ramp by default. On day one the effective cap is 20% of your campaign's daily follow limit, and it climbs linearly to 100% as the window completes. The cap never drops below 3 follows per day, so a warming account always makes some progress. The ramp anchors to the moment the account was connected (or to a warmup restart, if you re-enable it later).
Example: a campaign capped at 30 follows per day on a fresh account starts near 6 per day, passes the halfway point (about 18 per day) around day 7, and reaches the full 30 when the ramp ends.
A rule-of-thumb view in weeks. In the product, the ramp is proportional: 20% of your campaign cap rising to 100% across the warmup window, 14 days by default.
Each account has a skip control for people who know exactly what they are doing. The dashboard warns you when you use it: accounts that skip the ramp go straight to full daily limits, and new accounts are far more likely to be flagged that way. You can re-enable warmup at any time, which restarts the ramp.
Pacing that reads as human
Within your active-hours window, follows keep a minimum gap of roughly the window length divided by your daily cap, jittered between 0.6x and 1.3x so the cadence never turns metronomic. A 12-hour window with a cap of 30 averages one follow every 24 minutes or so, some closer together, some further apart. Windows can wrap midnight, and setting the start equal to the end means around the clock, though the daily cap still applies.
The ban-risk score
Every connected account carries a 0-100 risk score, recomputed continuously from its last 24 hours of behavior. Each contributing factor is shown in the dashboard with its own weight, so a rising score is never a mystery. Account health is simply the inverse: health = 100 minus risk.
| Level | Score | What happens |
|---|---|---|
| Low | 0-24 | Normal operation |
| Elevated | 25-49 | Worth a look; factors listed on the account page |
| High | 50-74 | An alert event fires the first time this level is crossed |
| Critical | 75-100 | Automatic safety hold: the account's campaigns pause |
| Risk factor | Trigger | Max weight |
|---|---|---|
| Rate limited | Hit Bluesky rate limits in the last 24h | 45 |
| Action errors | 3 or more failed actions in 24h | 25 |
| Overshoot | Followed past the safe daily pace | 20 |
| Error state | The account session is failing | 18 |
| New-account burst | Under 7 days old with more than 15 follows in a day | 15 |
| Very low follow-back | Under 5% across 20+ recent follows | 15 |
| Degraded health | Health score below 70 | 15 |
| Banned | Bluesky has banned the account | 100 |
Safety holds and getting out of them
When an account crosses into critical, SkyFollowing places a safety hold: its campaigns pause on their own, a pause event is recorded (and delivered to your webhooks if configured), and the account page lists exactly which factors drove the score. Other accounts in the workspace keep running. Once you have addressed the cause, the Resume control lifts the hold; risk keeps being recomputed, so an unresolved cause will simply hold it again.
Every scan also records a health snapshot, so the account page shows risk history over time rather than a single number out of context.
Frequently asked questions
Can I turn the safety systems off?
Warmup can be skipped per account, with a warning. Pacing, the risk score, and critical-level holds cannot be disabled; they are the difference between automation and liability.
Why is my effective daily cap lower than my campaign setting?
The account is inside its warmup window, so the effective cap is the ramp percentage applied to your campaign cap. The Accounts page shows warmup progress; the cap reaches your full setting when the ramp completes.
What should I do after a safety hold?
Read the factor list first. Rate limits or error bursts usually mean a session problem; reconnect fixes those. Overshoot or burst factors mean pacing was too hot; lower the cap before resuming.
Every setting on this page ships with safe defaults. Free for 7 days, no card required.