openai-auto-review-codex-2026
openai-auto-review-codex-2026
fields
| statementREQ | OpenAI will release Auto-review in Codex, a mode that replaces synchronous human approval at sandbox boundaries with review by a separate agent, enabling safer long-running agentic tasks without constant user oversight. | alignment.openai.com | 2026-05-03 |
| actorREQ | OpenAI | alignment.openai.com | 2026-05-03 |
| topic | AI safety / developer tooling | alignment.openai.com | 2026-05-03 |
| promised_at | 2026-04-30 | alignment.openai.com | 2026-05-03 |
| target_date | 2026-04-30 | alignment.openai.com | 2026-05-03 |
| shipped | true | alignment.openai.com | 2026-05-03 |
| shipped_at | 2026-04-23 | alignment.openai.com | 2026-05-03 |
| evidence_url | https://alignment.openai.com/auto-review/ | alignment.openai.com | 2026-05-03 |
history · 8 fields · 8 revisions
statement1 revision
OpenAI will release Auto-review in Codex, a mode that replaces synchronous human approval at sandbox boundaries with review by a separate agent, enabling safer long-running agentic tasks without constant user oversight.
current
Last week, we released Auto-review in Codex. Auto-review offers an alternative path. It replaces user approval at the sandbox boundary with review by a separate agent. Internally, Auto-review increases our confidence in running long agentic tasks without synchronous human oversight.
actor1 revision
OpenAI
current
Last week, we released Auto-review in Codex. Auto-review offers an alternative path. It replaces user approval at the sandbox boundary with review by a separate agent. Internally, Auto-review increases our confidence in running long agentic tasks without synchronous human oversight.
topic1 revision
AI safety / developer tooling
current
Last week, we released Auto-review in Codex. Auto-review offers an alternative path. It replaces user approval at the sandbox boundary with review by a separate agent. Internally, Auto-review increases our confidence in running long agentic tasks without synchronous human oversight.
promised_at1 revision
2026-04-30
current
Last week, we released Auto-review in Codex. Auto-review offers an alternative path. It replaces user approval at the sandbox boundary with review by a separate agent. Internally, Auto-review increases our confidence in running long agentic tasks without synchronous human oversight.
target_date1 revision
2026-04-30
current
Last week, we released Auto-review in Codex. Auto-review offers an alternative path. It replaces user approval at the sandbox boundary with review by a separate agent. Internally, Auto-review increases our confidence in running long agentic tasks without synchronous human oversight.
shipped1 revision
true
current
Last week, we released Auto-review in Codex. Auto-review offers an alternative path. It replaces user approval at the sandbox boundary with review by a separate agent. Internally, Auto-review increases our confidence in running long agentic tasks without synchronous human oversight.
shipped_at1 revision
2026-04-23
current
Last week, we released Auto-review in Codex. Auto-review offers an alternative path. It replaces user approval at the sandbox boundary with review by a separate agent. Internally, Auto-review increases our confidence in running long agentic tasks without synchronous human oversight.
evidence_url1 revision
https://alignment.openai.com/auto-review/
current
Last week, we released Auto-review in Codex. Auto-review offers an alternative path. It replaces user approval at the sandbox boundary with review by a separate agent. Internally, Auto-review increases our confidence in running long agentic tasks without synchronous human oversight.