Ask Pete (pre-flight checker)
Ask Pete is a lint + dry-run checker that tells you what would break in a sequence, blockscript, or scheduler campaign — before the night does. Point it at what you're about to run and it hands back a plain-language verdict and a list of findings, so you don't discover a dangling filter name or a target that never rises at 2am.
Pete is the simulated-operator persona Starship uses for its night simulations; the pre-flight checker is that same "would this actually run?" scrutiny wired into the editors.
Where it lives
A Check plan button sits in the Sequencer, Blockscript, and Schedule editors (it was labelled "Ask Pete" before v23.0.1; Pete still signs the report). Clicking it checks the current config — the sequence you're editing, the selected blockscript (even unsaved), or your live project list — and pops a report modal: a verdict pill (errors / warnings / all clear), each finding with a ✓ / ⚠ / ✗ icon and where it is, and Pete's one-line take ("I wouldn't run this yet — 3 things are broken" / "Looks clean, I'd hit start").
It's also an API: POST /api/pete/check {kind: "sequence" | "blockscript" | "scheduler", ...} returns the findings plus a friendly summary.
What it catches
Two layers:
- Static lint (reliable for "undefined") — every dangling reference:
- a filter that isn't in your wheel / library, - a slew with no coordinates and no target, - a run_sequence naming a plan that doesn't exist, - a blockscript goto to a state that isn't there, - an unknown action type, or a retry_connect for a device you don't have.
- Light dry-run — a sequence is compiled and walked (always finite) for a **time
estimate and value sanity; the scheduler is scanned against tonight's sky to flag a target that never clears your horizon (it would sit idle forever) or a malformed goal; a blockscript is checked for goto self-loops, an unparsable wait until time, and a session with no run_sequence / run_scheduler body** anywhere — it would idle in RUNNING until you force shutdown.
What it can't promise
Truly proving a script halts is the halting problem — no tool can. Pete flags the structural mistakes (a self-loop, a goal that can never complete) and is upfront that he can't certify termination. Treat a clean report as "no obvious breakage," not "guaranteed to finish."
How it fits with the simulators
Ask Pete is the quick, per-edit sanity check. For a fuller look before you trust a night, the Schedule page's Simulate tonight draws tonight's plan as a timeline (with per-target skip reasons). Use Ask Pete to catch mistakes as you build, and Simulate tonight to see how a valid plan actually plays out. (A multi-night campaign simulation exists in the scheduler's code but is not reachable from the app yet.) See the Sequencer, Blockscript, and Scheduler topics.