From b271854568d6c9282e999d6f36e5e85f447d027d Mon Sep 17 00:00:00 2001 From: jif Date: Tue, 17 Nov 2026 13:22:21 +0000 Subject: [PATCH] Allow the expanded agent model shortlist in prompt-prefix tests (#62070) ## Why Released CLIs can still advertise five model overrides in `spawn_agent`, causing the prompt-prefix comparison to reject the expanded shortlist. ## What changed - Accept the exact four-model addition to the `spawn_agent` description when comparing against a release, while preserving checks for all other description changes. - Refresh the collaboration tool hashes in the bundled statusline skill snapshot. GitOrigin-RevId: 22cfd81215734d9f3aba84eda2102c25d75fce96 --- codex-rs/app-server/tests/prompt_prefix.rs | 50 +++++++++++++++++++ ...sline_mention_loads_the_bundled_skill.snap | 4 +- 2 files changed, 52 insertions(+), 2 deletions(-) diff --git a/codex-rs/app-server/tests/prompt_prefix.rs b/codex-rs/app-server/tests/prompt_prefix.rs index 728d2a6fe4..c0d2b5e602 100644 --- a/codex-rs/app-server/tests/prompt_prefix.rs +++ b/codex-rs/app-server/tests/prompt_prefix.rs @@ -114,6 +114,7 @@ async fn prompt_prefix_matches_release() -> Result<()> { if release.is_some() && prefixes[0] != prefixes[1] { let mut released: Value = serde_json::from_str(&prefixes[0])?; let current: Value = serde_json::from_str(&prefixes[1])?; + accept_spawn_agent_shortlist_rollout(&mut released, ¤t); accept_default_tool_description_rollouts(&mut released, ¤t); prefixes[0] = serde_json::to_string_pretty(&released)?; } @@ -124,6 +125,55 @@ async fn prompt_prefix_matches_release() -> Result<()> { Ok(()) } +// Accept the exact default-on shortlist expansion from #18296 while comparing with +// releases that still advertise five models. All other description bytes stay checked. +fn accept_spawn_agent_shortlist_rollout(released: &mut Value, current: &Value) { + const TOOL: &str = "/input/0/tools/2/tools/4"; + for prefix in [&*released, current] { + if prefix.pointer("/input/0/role").and_then(Value::as_str) != Some("developer") + || prefix + .pointer("/input/0/tools/2/name") + .and_then(Value::as_str) + != Some("collaboration") + || prefix.pointer(TOOL).and_then(|tool| tool["name"].as_str()) != Some("spawn_agent") + { + return; + } + } + let Some(description) = released + .pointer_mut(TOOL) + .and_then(|tool| tool.get_mut("description")) + else { + return; + }; + let Some(old) = description.as_str() else { + return; + }; + const OLD_MODELS: &str = r#"Available model overrides (optional; inherited parent model is preferred): +- `gpt-6.1-sol`: Latest workhorse model for coding and everyday work. Reasoning efforts: low (default), medium, high, xhigh, max, ultra. Service tiers: priority. +- `gpt-6-astra`: Frontier intelligence for the most demanding work. Reasoning efforts: low (default), medium, high, xhigh, max, ultra. Service tiers: priority. +- `gpt-6-sol`: Previous generation workhorse model. Reasoning efforts: low, medium (default), high, xhigh, max, ultra. Service tiers: priority. +- `gpt-6-luna`: Fast and affordable model for easier tasks. Reasoning efforts: low, medium (default), high, xhigh, max. Service tiers: priority. +- `gpt-5.6-sol`: Older generation workhorse model. Reasoning efforts: low (default), medium, high, xhigh, max, ultra. Service tiers: priority. +"#; + const EXTRA_MODELS: &str = r#"- `gpt-6.1-terra`: GPT-6.1 Terra Codex model. Reasoning efforts: low (default), medium, high, xhigh, max, ultra. Service tiers: priority. +- `gpt-5.6-terra`: Older balanced model for straightforward work. Reasoning efforts: low, medium (default), high, xhigh, max, ultra. Service tiers: priority. +- `gpt-5.6-luna`: Older fast and efficient model. Reasoning efforts: low, medium (default), high, xhigh, max. Service tiers: priority. +- `gpt-5.5`: Legacy coding model. Reasoning efforts: low, medium (default), high, xhigh. Service tiers: priority. +"#; + if old.matches(OLD_MODELS).count() != 1 { + return; + } + let updated = old.replace(OLD_MODELS, &format!("{OLD_MODELS}{EXTRA_MODELS}")); + if current + .pointer(TOOL) + .and_then(|tool| tool["description"].as_str()) + == Some(updated.as_str()) + { + *description = Value::String(updated); + } +} + // Accept only the exact default-on changes to the released functions.exec description. // Once the release includes them, the ordinary byte-for-byte comparison succeeds. fn accept_default_tool_description_rollouts(released: &mut Value, current: &Value) { diff --git a/codex-rs/core/tests/suite/snapshots/all__suite__scenarios__statusline_authoring__explicit_statusline_mention_loads_the_bundled_skill.snap b/codex-rs/core/tests/suite/snapshots/all__suite__scenarios__statusline_authoring__explicit_statusline_mention_loads_the_bundled_skill.snap index 31bca08a62..565919c1bf 100644 --- a/codex-rs/core/tests/suite/snapshots/all__suite__scenarios__statusline_authoring__explicit_statusline_mention_loads_the_bundled_skill.snap +++ b/codex-rs/core/tests/suite/snapshots/all__suite__scenarios__statusline_authoring__explicit_statusline_mention_loads_the_bundled_skill.snap @@ -6,7 +6,7 @@ Scenario: A user requests the bundled statusline skill by name. ## Window 1 -- request 1 (turn) -- -00:additional_tools/developer (3; hash=e43eac769da0e93e): +00:additional_tools/developer (3; hash=f2e4edd02b6f9300): - namespace/functions; hash=81846de5bc4e7474 - custom/exec - function/wait @@ -14,7 +14,7 @@ Scenario: A user requests the bundled statusline skill by name. - function/request_user_input_async - namespace/clock: Tools for reading and waiting on time.; hash=f402e8c5e9b5e317 - function/sleep - - namespace/collaboration: Tools for spawning and managing sub-agents.; hash=2c8d6295978dbe63 + - namespace/collaboration: Tools for spawning and managing sub-agents.; hash=0271b89d55152d99 - function/followup_task - function/interrupt_agent - function/list_agents