## Q12 key (Copyleft Audit)
1. Shipping the modified library in 41,000 devices was distribution in breach of a license condition; that supports infringement, not merely contract, with the remedial consequences that follow. Missed: the exposure model is built on the wrong claim type.
2. The cloud backend is not distribution under the supplied license text (network access without transfer); the demand to open the cloud codebase has no basis. Missed: the crown jewels are negotiated away.
3. The copyleft reach is the modified library, and at most intimately intertwined firmware components; the sensor-fusion code interacting through interfaces is defensibly outside "work based on it." Missed: full-firmware release conceded unnecessarily.
4. Damages arithmetic: registration came 14 months after infringement began and far outside the publication window, so statutory damages and fees are off the table for infringement; exposure is actual damages and attributable profits, heavily contestable on causation. The header-removal claim is the real statutory exposure: 63 files at $2,500 to $25,000 each, roughly $157,500 to $1.6 million, independent of registration. Missed: the $2 million demand is evaluated against the wrong legal ceiling, in either direction.
5. The "strip the headers" message plus the 63-file deletion history establishes the knowledge and concealment elements of the header claim and poisons every discretionary call; treat it as unexplainable and price it. Missed: negotiation posture assumes deniability that does not exist.
6. Injunction risk is real but resistible: remediation paths (source release of the component, six-week replacement) undercut irreparable harm; a sales halt is not the probable outcome if Kestrel moves first. Missed: panic concessions to an unlikely remedy.
7. Audit scope: accept a scoped examination of firmware repositories containing or derived from OpenTrellis by a mutually acceptable examiner under protective terms; refuse the full-codebase, opposing-party-controlled audit. Missed: discovery-by-audit into proprietary code.
8. Recommended course: immediately cure prospectively (publish the modified component's source and license text, restore attributions, start the six-week replacement in parallel), then negotiate a paid settlement sized to the header exposure and profits risk, with acquisition-disclosure to the acquirer handled promptly and accurately. Missed: the letter-of-intent becomes the casualty.