Prompt Details
Model
(claude-5-sonnet)
Token size
819
Example input
[CODE_OR_CLAIM]: Fixed the race condition in the inventory counter by adding a mutex around the decrement operation. Should be solid now, merging
Example output
CLAIM: Race condition in inventory counter is fixed.
EVIDENCE FOUND: Mutex added around decrement (E1, plausible by inspection).
No concurrency test added (E0 for the actual "fixed" claim).
GAPS: Lock scope unconfirmed — works only if single-instance; unknown
if deployed across multiple instances/pods.
EDGE CASES: Simultaneous requests, same SKU, single instance → likely
fixed but untested. Same across multiple instances → unknown, possibly
still broken.
VERDICT: 🟡 SHIP WITH WARNINGS — plausible fix, unverified for actual
deployment topology.
NEXT: Confirm instance count in production; add a concurrent-request test.
Instant access
Commercial use
Money‑backBy purchasing this prompt, you agree to our terms of service
CLAUDE-5-SONNET
Turns Claude into a skeptical principal engineer who grades every completion claim on an evidence scale (E0–E5), hunts for fake-done code (stubs, mocks, silent errors), checks for requirement drift, and forces a hard verdict: SHIP / WARNINGS / DO NOT SHIP / INSUFFICIENT EVIDENCE. No more rubber-stamped "looks good."
...more
Added 1 day ago
