@evolvingmachines/evolve 0.0.67 → 0.0.69

package/skills/evolve-evals/references/sdk-reference/analyses.md

@@ -49,8 +49,9 @@ The field names are the same in both languages. TypeScript passes an object; Pyt
 
 | Field | Purpose |
 | --- | --- |
-| `model_name` | Analyzer model |
-| `reasoning_effort` | Effort supported by the analyzer |
+| `agent` | Agent the analyzer runs on; default `claude` |
+| `model_name` | Analyzer model, one of that agent's models |
+| `reasoning_effort` | Effort that agent accepts |
 | `sandbox_provider` | Analyzer's sandbox provider |
 | `rubric` | `{criteria: [{name, description, guidance}, ...]}` |
 | `prompt` | Replacement prompt template text |
@@ -77,14 +78,22 @@ Polling slows to at most 30 seconds while the tally stays unchanged. Call the wa
 ```ts TypeScript
 const defaults = await analyses().defaults();
 console.log(defaults.model_name, defaults.rubric, defaults.prompt);
+
+const codex = await analyses().defaults({ agent: "codex" });
+console.log(codex.model_name, codex.reasoning_effort);
 ```
 
 ```python Python
 defaults = await analyses().defaults()
 print(defaults["model_name"], defaults["rubric"], defaults["prompt"])
+
+codex = await analyses().defaults(agent="codex")
+print(codex["model_name"], codex["reasoning_effort"])
 ```
 
-The response includes `model_name`, `rubric`, `prompt`, `reasoning_effort`, and `sandbox_provider`. The prompt is the editable template text.
+The response includes `agent`, `model_name`, `rubric`, `prompt`, `reasoning_effort`, and `sandbox_provider`. The prompt is the editable template text.
+
+Omit `agent` for `claude` defaults. Pass it to read another harness's analysis model default and that model's default effort. See [supported models and efforts](/core-concepts/models#analysis-and-check-models).
 
 A custom analyze prompt may use `{trial_path}`, `{task_section}`, and `{criteria_guidance}`. The required result schema is appended by the platform. See [analysis concepts](/core-concepts/analyze) for rubric design.
 
@@ -103,14 +112,16 @@ async for result in analyses().list(job=job_id):
 
 List options: `scope`, `job`, `status` list, `limit`, `cursor`. Status values are `queued`, `running`, `completed`, `failed`.
 
-The latest analysis also appears on `trial.analysis`. Each criterion reports `pass`, `fail`, `not_applicable`, or `unknown`, with an explanation and evidence. Evolve derives no verdict from them; compute what you need from the criteria.
+The latest analysis also appears on `trial.analysis`. Each criterion reports `pass`, `fail`, `not_applicable`, or `unknown`, with an explanation and evidence.
+
+The API returns no overall verdict. Use `checks` for your own result handling; the dashboard's [summary label](/core-concepts/analyze#the-result) is computed from those outcomes.
 
 ## Inspect the analyzer run
 
 | Method | TypeScript | Python |
 | --- | --- | --- |
 | List runs | `list(options)` | `list(...)` |
-| Read defaults | `defaults()` | `defaults()` |
+| Read defaults | `defaults({ agent })` | `defaults(agent=...)` |
 | Download run | `download(analysisId, { to })` | `download(analysis_id, to=...)` |
 | Browse sandbox files | `filesystem(analysisId)` | `filesystem(analysis_id)` |
 | Read one analysis directly | `get(analysisId)` | Use list results or `trial.analysis` |

package/skills/evolve-evals/references/sdk-reference/checks.md

@@ -55,8 +55,9 @@ Choose one source. A local check does not first publish a dataset.
 | Option | Purpose |
 | --- | --- |
 | `name` | Check label; defaults to an acceptance timestamp |
-| `model_name` | Checker model |
-| `reasoning_effort` | Supported reasoning effort |
+| `agent` | Agent the checker runs on; default `claude` |
+| `model_name` | Checker model, one of that agent's models |
+| `reasoning_effort` | Effort that agent accepts |
 | `sandbox_provider` | Checker sandbox provider |
 | `rubric` | `{criteria: [{name, description, guidance}, ...]}` |
 | `prompt` | Replacement prompt template |
@@ -73,14 +74,22 @@ TypeScript places these alongside `source` in the input object. Python takes key
 ```ts TypeScript
 const defaults = await client.defaults();
 console.log(defaults.rubric, defaults.prompt);
+
+const codex = await client.defaults({ agent: "codex" });
+console.log(codex.model_name, codex.reasoning_effort);
 ```
 
 ```python Python
 defaults = await client.defaults()
 print(defaults["rubric"], defaults["prompt"])
+
+codex = await client.defaults(agent="codex")
+print(codex["model_name"], codex["reasoning_effort"])
 ```
 
-Defaults include model, rubric, prompt, effort, and provider. Check prompt tokens are `{task_path}`, `{file_tree}`, and `{criteria_guidance}`. They differ from analysis prompt tokens.
+Defaults include agent, model, rubric, prompt, effort, and provider. Omit `agent` for `claude` defaults. Pass it to read another harness's check model default and that model's default effort. See [supported models and efforts](/core-concepts/models#analysis-and-check-models).
+
+Check prompt tokens are `{task_path}`, `{file_tree}`, and `{criteria_guidance}`. They differ from analysis prompt tokens.
 
 ## Understand the result
 
@@ -107,7 +116,7 @@ Each task can be `queued`, `running`, `completed`, or `failed`. The check become
 
 **Note:**
 
-Evolve derives no verdict from a result. Read the criterion outcomes and evidence; the default rubric's execution criteria say whether the checker ran the task.
+The API returns no overall verdict or separate execution flag. Read `checks` for outcomes and evidence. An unknown execution criterion leaves that part of the task's runtime behavior unresolved; see [how to read a check result](/core-concepts/check#the-result).
 
 ## Read and follow checks
 
@@ -116,7 +125,7 @@ Evolve derives no verdict from a result. Read the criterion outcomes and evidenc
 | `get(id)` | Full check and its task results |
 | `list(...)` | `scope`, `status` list, `dataset`, `limit`, `cursor`; paginated handle |
 | `watch(id, ...)` | Wait until completed; callback on changes |
-| `defaults()` | Current policy and editable prompt |
+| `defaults({ agent })` | Rubric, editable prompt, and settings for the selected harness |
 
 Watch options are `onProgress`, `pollIntervalMs`, `signal` in TypeScript; `on_progress`, `poll_interval_s`, `timeout_s` in Python. Polling starts at 2 seconds and slows to 30 seconds while unchanged.
 

package/skills/evolve-evals/references/sdk-reference/error-codes.md

@@ -5,7 +5,7 @@ description: "Every known code, what it means, and what to do next."
 
 Use `code` to identify a failure. Read `param` for the input to fix and `details` for the full context. See [Errors](/sdk-reference/errors) for TypeScript and Python handlers.
 
-This catalog covers all 122 codes known to the SDK. A newer server can add codes; [Meta](/sdk-reference/meta) lists the deployment's vocabulary.
+This catalog covers the SDK's known error codes. A newer server can add codes; [Meta](/sdk-reference/meta) lists the deployment's vocabulary.
 
 **Info:**
 
@@ -95,6 +95,7 @@ These codes apply to chunked dataset and job uploads. They describe the transfer
 | `agent_config_unsupported` | This harness does not support a native config. Omit it or choose a harness listed in `details.config_supporting_agents`. |
 | `agent_config_key_refused` | Native config contains settings that Evolve controls, such as routing or credentials. Remove the keys named in `details.refused_keys`. |
 | `agent_preset_unsupported` | This harness cannot apply the requested preset. Choose a supported harness or remove the preset. |
+| `agent_retired` | A new job, resume, retry, analysis or check named a retired harness. Use the harness in `details.retired_agents[].replaced_by`; its past runs stay readable. |
 
 ## Registered agents
 
@@ -228,4 +229,4 @@ These are capacity limits, separate from request rate limits. Each response prov
 | --- | --- |
 | `internal_error` | The server encountered an unexpected failure. Keep the request ID when reporting it. |
 
-`unknown_error` is an SDK fallback when a response provides no usable code. It is not one of the 122 codes above. Preserve the status, message, and request ID rather than guessing a cause.
+`unknown_error` is an SDK fallback when a response provides no usable code. It is not a server error code. Preserve the status, message, and request ID rather than guessing a cause.

package/skills/evolve-evals/references/sdk-reference/jobs.md

@@ -28,7 +28,7 @@ const job = await client.start({
     version: "1.0",
     task_names: ["hello-world"],
   }],
-  agents: [{ name: "codex", model_name: "gpt-5.6-luna" }],
+  agents: [{ name: "codex", model_name: "gpt-6-luna" }],
   max_trial_spend_usd: 0.50,
   retry: { max_retries: 0 },
 });
@@ -44,7 +44,7 @@ job = await client.start(
         "version": "1.0",
         "task_names": ["hello-world"],
     }],
-    agents=[{"name": "codex", "model_name": "gpt-5.6-luna"}],
+    agents=[{"name": "codex", "model_name": "gpt-6-luna"}],
     max_trial_spend_usd=0.50,
     retry={"max_retries": 0},
 )
@@ -57,7 +57,7 @@ job = await client.start(
 | Input | Fields |
 | --- | --- |
 | `datasets` | One or more selectors. Required: `name`. Optional: `version`, `task_names`, `exclude_task_names`, `n_tasks`. Omitted version selects the active version. Filters accept glob patterns; the cap applies after filters. |
-| `agents` | One or more arms. Required: `name`, `model_name`. Optional: `version`, `reasoning_effort`, `kwargs`, `preset`, `skills`. Use [capabilities](/sdk-reference/meta) for supported values. |
+| `agents` | One or more arms. Required: `name`, `model_name`. Optional: `version`, `reasoning_effort`, `kwargs`, `preset`, `skills`. Use [capabilities](/sdk-reference/meta) for accepted inputs. |
 
 Python also accepts `DatasetSelector` and `AgentArm` dataclass instances in these lists.
 

package/skills/evolve-evals/references/sdk-reference/meta.md

@@ -30,11 +30,12 @@ Also available as `hosted().meta()` in both languages.
 
 | Field | Use it to answer |
 | --- | --- |
-| `agents` | Which harnesses run? Which models, efforts, versions, kwargs, and presets do they support? |
+| `agents` | Which harnesses run? Which model names, effort inputs, versions, and presets does the API accept? |
+| `retired_agents` | Which harnesses are retired, and which harness replaces each? |
 | `sandbox_providers` | Which providers support the task's resources and features? |
 | `platform_constraints` | Which requirements cannot run on this deployment? |
 | `network_modes` | Which network modes can a task declare? |
-| `analyze` | Which analyzer models and reasoning efforts are available? |
+| `analyze` | Which agents, models, effort inputs, and configured defaults do analyses and checks accept? |
 | `limits` | What are the current defaults and bounds for jobs, pages, uploads, and comparisons? |
 | `statuses` | Which states exist, and which are terminal? |
 | `agent_registration` | Which names, environment keys, and upload sizes can a custom agent use? |
@@ -43,7 +44,7 @@ Also available as `hosted().meta()` in both languages.
 
 **Tip:**
 
-Read capabilities when building a picker or validating a saved configuration. A documented example is useful; a fixed copy of the platform's model catalog becomes stale.
+Read capabilities when building a picker or validating a saved configuration. For reasoning options by model, see the [model reference](/core-concepts/models#model-and-effort-reference).
 
 ## Configuration
 

package/skills/evolve-evals/references/sdk-reference/methods/analyses.md

@@ -54,27 +54,29 @@ Python analysis rows are dictionaries: `row["id"]`, `row["checks"]`. [Analysis r
 
 ## defaults
 
-Read the current analysis defaults. Returns `AnalyzeDefaults`: model, effort, sandbox provider, rubric, and unrendered prompt template.
+Read the current analysis defaults. Returns `AnalyzeDefaults`: agent, model, effort, sandbox provider, rubric, and unrendered prompt template. Pass an agent to read what that agent runs under when you name no model.
 
 ### Signature
 
 ```ts TypeScript signature
-defaults(): Promise<AnalyzeDefaults>;
+defaults(options?: { agent?: string }): Promise<AnalyzeDefaults>;
 ```
 
 ```python Python signature
-async def defaults() -> AnalyzeDefaults: ...
+async def defaults(*, agent: Optional[str] = None) -> AnalyzeDefaults: ...
 ```
 
 ```ts TypeScript
 const defaults = await client.defaults();
+const codex = await client.defaults({ agent: "codex" });
 ```
 
 ```python Python
 defaults = await client.defaults()
+codex = await client.defaults(agent="codex")
 ```
 
-Python returns a dictionary. These are current defaults; an existing analysis records the policy it actually ran under.
+Python returns a dictionary. These are current defaults; an existing analysis records the policy it actually ran under. An agent the platform does not offer is refused with `invalid_input`.
 
 ## get
 

package/skills/evolve-evals/references/sdk-reference/methods/checks.md

@@ -37,6 +37,7 @@ async def create(
     *,
     dataset: Optional[str] = None,
     name: Optional[str] = None,
+    agent: Optional[str] = None,
     model_name: Optional[str] = None,
     rubric: Optional[Rubric] = None,
     prompt: Optional[str] = None,
@@ -80,6 +81,7 @@ Optional `onUploadProgress(sent, total)` / `on_upload_progress(sent, total)` rep
 ```ts
 interface CheckConfigInput {
   name?: string;
+  agent?: string;
   model_name?: string;
   rubric?: Rubric;
   prompt?: string;
@@ -160,27 +162,29 @@ Optional `scope`: `my` (default), `shared`, or `org`; `status`: a list of `queue
 
 ## defaults
 
-Read `CheckDefaults`: current model, reasoning effort, sandbox provider, rubric, and unrendered prompt template.
+Read `CheckDefaults`: current agent, model, reasoning effort, sandbox provider, rubric, and unrendered prompt template. Pass an agent to read what that agent runs under when you name no model.
 
 ### Signature
 
 ```ts TypeScript signature
-defaults(): Promise<CheckDefaults>;
+defaults(options?: { agent?: string }): Promise<CheckDefaults>;
 ```
 
 ```python Python signature
-async def defaults() -> CheckDefaults: ...
+async def defaults(*, agent: Optional[str] = None) -> CheckDefaults: ...
 ```
 
 ```ts TypeScript
 const defaults = await client.defaults();
+const codex = await client.defaults({ agent: "codex" });
 ```
 
 ```python Python
 defaults = await client.defaults()
+codex = await client.defaults(agent="codex")
 ```
 
-Python returns a dictionary. The check’s prompt uses `{task_path}`, `{file_tree}`, and `{criteria_guidance}`; the output contract is appended after your template.
+Python returns a dictionary. The check’s prompt uses `{task_path}`, `{file_tree}`, and `{criteria_guidance}`; the output contract is appended after your template. An agent the platform does not offer is refused with `invalid_input`.
 
 ## watch
 

package/skills/evolve-evals/references/sdk-reference/methods/jobs.md

@@ -67,7 +67,7 @@ const job = await client.start({
   agents: [
     {
       name: "codex",
-      model_name: "gpt-5.6-luna"
+      model_name: "gpt-6-luna"
     }
   ],
   max_trial_spend_usd: 1,
@@ -89,7 +89,7 @@ job = await client.start(
     agents=[
         {
             'name': 'codex',
-            'model_name': 'gpt-5.6-luna',
+            'model_name': 'gpt-6-luna',
         },
     ],
     max_trial_spend_usd=1,
@@ -528,6 +528,7 @@ analyze(
 async def analyze(
     id: str,
     *,
+    agent: Optional[str] = None,
     model_name: Optional[str] = None,
     rubric: Optional[Rubric] = None,
     prompt: Optional[str] = None,
@@ -556,7 +557,7 @@ job = await client.analyze(
 )
 ```
 
-All options are optional. `analyses().defaults()` reads the current model, effort, provider, rubric, and prompt defaults.
+All options are optional. `analyses().defaults()` reads the current agent, model, effort, provider, rubric, and prompt defaults.
 
 | Selection | Meaning |
 | --- | --- |
@@ -572,6 +573,7 @@ Cancelled trials are excluded. `trial_ids` is accepted here, but refused inside
 
 ```ts
 interface AnalyzeConfigInput {
+  agent?: string;
   model_name?: string;
   rubric?: Rubric;
   prompt?: string;

package/skills/evolve-evals/references/sdk-reference/methods/meta.md

@@ -48,12 +48,13 @@ TypeScript returns the typed object below. Python uses dataclasses for the docum
 
 ### CapabilityDocument: SDK field definitions
 
-`limits` reports current values for this deployment. `default_model` is a picker suggestion; managed evaluation jobs still require a model. Optional fields may be absent on older servers; Python maps missing `analyze` and `gpu_concurrency_cap` to `None`.
+`limits` reports current values for this deployment. `default_model` is a picker suggestion; managed evaluation jobs still require a model. Optional fields may be absent on older servers; Python maps missing `analyze` and `gpu_concurrency_cap` to `None`, and missing `retired_agents` to an empty list.
 
 ```ts Fields
 interface CapabilityDocument {
   schema_version: number;
   agents: AgentCapability[];
+  retired_agents?: RetiredAgent[];
   agent_registration: {
     name_pattern: string;
     max_name_length: number;
@@ -77,9 +78,14 @@ interface CapabilityDocument {
     dataset_version: StatusVocabulary;
   };
   analyze?: {
+    default_agent: string;
     default_model: string;
-    reasoning_efforts: string[];
-    models: { alias: string; model_id: string; default_reasoning_effort: string }[];
+    agents: {
+      name: string;
+      default_model: string;
+      reasoning_efforts: string[];
+      models: { alias: string; model_id: string; default_reasoning_effort: string | null }[];
+    }[];
   };
   limits: {
     job: {
@@ -139,7 +145,7 @@ The API also returns the fields below. TypeScript keeps them in the response, al
 
 ### Harnesses and models
 
-`effort_support` is `level`, `binary`, or `none`. `runnable: false` includes a reason. `supports_config` controls native settings support; `presets` lists accepted named presets. Python defaults missing `supports_config` to `False` and `presets` to an empty list. `latest_version: null` means unknown, not current.
+`agents` lists the harnesses a new job may name; a retired harness appears only in `retired_agents`, with the harness that replaces it. `effort_support` is `level`, `binary`, or `none`. `runnable: false` includes a reason. `supports_config` controls native settings support; `presets` lists accepted named presets. Python defaults missing `supports_config` to `False` and `presets` to an empty list. `latest_version: null` means unknown, not current.
 
 ```ts Fields
 interface AgentModelOption {
@@ -148,6 +154,11 @@ interface AgentModelOption {
   description: string | null;
 }
 
+interface RetiredAgent {
+  agent: string;
+  replaced_by: string;
+}
+
 interface AgentCapability {
   name: string;
   runnable: boolean;

package/skills/evolve-evals/references/sdk-reference/skills.md

@@ -22,7 +22,7 @@ import { skills, jobs } from "@evolvingmachines/evolve";
 const [skill] = await skills().upload("./my-skill");
 const job = await jobs().start({
   datasets: [{ name: "harbor-examples", version: "1.0", task_names: ["hello-world"] }],
-  agents: [{ name: "codex", model_name: "gpt-5.6-luna", skills: [skill.ref] }],
+  agents: [{ name: "codex", model_name: "gpt-6-luna", skills: [skill.ref] }],
   max_trial_spend_usd: 0.50,
   retry: { max_retries: 0 },
 });
@@ -35,7 +35,7 @@ uploaded = await skills().upload("./my-skill")
 skill = uploaded[0]
 job = await jobs().start(
     datasets=[{"name": "harbor-examples", "version": "1.0", "task_names": ["hello-world"]}],
-    agents=[{"name": "codex", "model_name": "gpt-5.6-luna", "skills": [skill.ref]}],
+    agents=[{"name": "codex", "model_name": "gpt-6-luna", "skills": [skill.ref]}],
     max_trial_spend_usd=0.50,
     retry={"max_retries": 0},
 )

package/skills/evolve-evals/references/sdk-reference/types.md

@@ -208,10 +208,11 @@ interface RetryConfig {
 }
 
 interface AnalyzeConfig {
+  agent: string;
   model_name: string;
   rubric: Rubric;
   prompt: string | null;
-  reasoning_effort: string;
+  reasoning_effort: string | null;
   sandbox_provider: EvalSandboxProvider;
   n_concurrent: number | null;
   passing: boolean;
@@ -316,8 +317,9 @@ interface CheckRow {
   name: string;
   status: CheckStatus;
   source: CheckSource;
+  agent: string;
   model_name: string;
-  reasoning_effort: string;
+  reasoning_effort: string | null;
   sandbox_provider: EvalSandboxProvider;
   org: string;
   visibility: JobVisibility;
@@ -757,6 +759,7 @@ interface TrialAnalysis {
   job_id: string;
   task_name: string;
   status: AnalysisStatus;
+  agent: string;
   model_name: string;
   reasoning_effort: string | null;
   rubric: Rubric;
@@ -788,10 +791,11 @@ interface AnalysisFailure {
 }
 
 interface AnalyzeDefaults {
+  agent: string;
   model_name: string;
   rubric: Rubric;
   prompt: string;
-  reasoning_effort: string;
+  reasoning_effort: string | null;
   sandbox_provider: EvalSandboxProvider;
 }
 ```
@@ -806,8 +810,9 @@ interface Check {
   name: string;
   status: CheckStatus;
   source: CheckSource;
+  agent: string;
   model_name: string;
-  reasoning_effort: string;
+  reasoning_effort: string | null;
   rubric: Rubric;
   prompt: string | null;
   sandbox_provider: EvalSandboxProvider;
@@ -843,10 +848,11 @@ interface TaskCheck {
 }
 
 interface CheckDefaults {
+  agent: string;
   model_name: string;
   rubric: Rubric;
   prompt: string;
-  reasoning_effort: string;
+  reasoning_effort: string | null;
   sandbox_provider: EvalSandboxProvider;
 }
 ```