Answer: Remove the retired `untrusted` approval policy from user or project configuration, profile files, startup scripts, and managed defaults. For interactive, read-only use, set `sandbox_mode = "read-only"` and `approval_policy = "on-request"`, or run `codex --sandbox read-only --ask-for-approval on-request`. Status: Confirmed Entry: tool/codex/permissions-and-security/migrate-untrusted-approval-policy Revision: 2 Digest: sha256:83152cd59a663ba9 (changes whenever the content changes) Timeline: Created: 2026-11-13T03:47:02Z Last changed: 2026-11-19T13:25:16Z (trust level: — → untrusted) Last verified: 2026-11-20T13:34:39Z Next check: 2026-11-21T13:34:39Z Check interval: Daily Expected change: No fixed schedule Data: Field | Value | Source | As of retired setting | untrusted | [1] | 2026-11-20 sandbox mode | read-only | [1] | 2026-11-20 approval policy | on-request | [1] | 2026-11-20 CLI command | codex --sandbox read-only --ask-for-approval on-request | [1] | 2026-11-20 config file | ~/.codex/config.toml | [1] | 2026-11-20 trust level | untrusted | [1] | 2026-11-20 Sources: [1] OpenAI. "Agent approvals & security". https://learn.chatgpt.com/docs/agent-approvals-security.md Role: primary. Accessed: 2026-11-20T13:34:39Z Quote: "Codex and ChatGPT Work no longer support `approval_policy = "untrusted"`." (Supports: retired setting) Quote: "sandbox_mode = "read-only"" (Supports: sandbox mode) Quote: "approval_policy = "on-request"" (Supports: approval policy) Quote: "Or run `codex --sandbox read-only --ask-for-approval on-request`." (Supports: CLI command) Quote: "`~/.codex/config.toml`" (Supports: config file) Quote: "trust_level = "untrusted"" (Supports: trust level) Snapshot SHA-256: fa9553edc8b520d62f0c0cf6a26382347beab09b9ff872ae04409fa3bd97a257 (captured 2026-11-18T13:28:47Z) Verification: Method: Automated extraction from [1] with deepseek-flash (JSON output, thinking disabled). Every value was checked by code to appear verbatim in a quoted passage of [1]. On this check the text of [1] was unchanged since 2026-11-18T13:28:47Z, so the values verified then were kept. Cross-check: not-available at 2026-11-20T13:34:39Z. No independent second source is configured for this entry Last successful check: 2026-11-20T13:34:39Z Last attempt: 2026-11-20T13:34:39Z Notes and limitations: Extracted automatically from a single official source (section "Migrate from the retired `untrusted` approval policy"). Related: https://latestn.com/tool/codex/permissions-and-security/permission-modes-overview What permission modes does ChatGPT and Codex offer for local actions? https://latestn.com/tool/codex/permissions-and-security/prepare-evidence-for-hardening What evidence should I provide to Codex Security when proposing security hardening? https://latestn.com/tool/codex/permissions-and-security/prepare-evidence-vulnerability-report What evidence should I provide to Codex before writing a vulnerability report? https://latestn.com/tool/codex/permissions-and-security/manage-privileges How do I manage privileges and permissions for the Codex GitHub Action? https://latestn.com/tool/codex/permissions-and-security/manual-security-review-of-changes How do I run a manual security review of code changes in Codex? https://latestn.com/tool/codex/permissions-and-security/migrate-sandbox-to-profiles How do I migrate Codex from older sandbox settings to permission profiles? https://latestn.com/tool/ All entries in AI tools Change log: Time | Event | Change | Source 2026-11-19T13:25:16Z | changed | trust level: — → untrusted | [1] 2026-11-19T13:25:16Z | changed | config file: — → ~/.codex/config.toml | [1] 2026-11-19T13:25:16Z | changed | config file path: ~/.codex/config.toml → — | [1] 2026-11-19T13:25:16Z | changed | project trust level: untrusted → — | [1] 2026-11-19T13:25:16Z | changed | retired setting: approval_policy = "untrusted" → untrusted | [1] 2026-11-13T03:47:02Z | added | Entry created | — Changes for this entry as a feed: https://latestn.com/tool/codex/permissions-and-security/migrate-untrusted-approval-policy/latest Challenge this entry: Open challenges: None How: If a value here is wrong or out of date, send what is wrong and a link to evidence. Challenges are queued and checked against the evidence; they do not change the entry by themselves. If the entry changes as a result, the change log says so. Form: https://latestn.com/dispute/tool/codex/permissions-and-security/migrate-untrusted-approval-policy POST: https://latestn.com/dispute/tool/codex/permissions-and-security/migrate-untrusted-approval-policy with claim, evidence_url, and optionally field and contact (form-encoded or JSON) GET: https://latestn.com/dispute/tool/codex/permissions-and-security/migrate-untrusted-approval-policy?claim={what is wrong}&evidence_url={link} How to cite: LatestN. (2026). How do I migrate from the retired untrusted approval policy in Codex? Revision 2, last verified 2026-11-20T13:34:39Z. https://latestn.com/tool/codex/permissions-and-security/migrate-untrusted-approval-policy (accessed 2026-11-21T10:07:37Z). @misc{LatestN_tool_codex_permissions_and_security_migrate_untrusted_approval_policy_r2, title = {How do I migrate from the retired untrusted approval policy in Codex?}, author = {{LatestN}}, year = {2026}, howpublished = {\url{https://latestn.com/tool/codex/permissions-and-security/migrate-untrusted-approval-policy}}, note = {Revision 2; digest sha256:83152cd59a663ba9; last verified 2026-11-20T13:34:39Z; accessed 2026-11-21T10:07:37Z} } Reuse: License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/) Attribution: Copying and adapting are allowed, including commercially, if you credit LatestN and link to the entry you used. Each entry ends with a ready-made citation. URL: https://latestn.com/tool/codex/permissions-and-security/migrate-untrusted-approval-policy