I fixed 30 of 31 failures. Nineteen were left.
A checker that stops at the first problem per item reports a queue depth, not a defect count.
We use Google Analytics to see which pages help people. No ads, no profiles. Privacy Policy
Driving AI agents to software that is verified on a real device, not just compiled.
39 posts
A checker that stops at the first problem per item reports a queue depth, not a defect count.
A named failure is preventable. "The AI messed up" is not.
A verification that asserts on the surface you wrote to will agree with you every time.
A budget expressed as a value can be read, derived from and audited. Buried in a conditional it can only be obeyed or removed, and it gets removed.
A command's promises decay silently through ordinary edits. Per-command assertions plus a fixture that removes one guarantee is what makes the decay loud.
Tier selection is not a preference, it is a function of three facts about the task. Two platform constraints rule out most of the combinations you would guess.
An always-loaded rulebook competes with itself for attention. Matching the tool target and injecting only what fires turns 634 rules into the twelve that apply.
Parallel agents sharing a working tree corrupt each other silently. Worktrees fix it, and cost 4.3 seconds and 364 MB each — not the 200ms the contract claims.
An agent reporting success has usually verified the call returned, not the artifact. Reading its own output back is the check it skips most often.
Workers report what they did; nothing reports what nobody did. A pass whose only question is what is missing finds the gaps no individual report can.
Whether an automated repair may apply unattended is arithmetic, not judgement: how certain the fix is against how much it can damage if wrong.
An autonomous drive given a backlog will reach for the interesting item. A head lock forces each item to be dispatched or explicitly dequeued, never skipped.