main.js · part 308

60 text occurrences from desktop/Cursor.app/Contents/Resources/app/extensions/cursor-agent-host/dist/main.js, part 308. Every entry preserves the shipped literal and its saved verdict or selection reason.

File contents and all parts · All files

${t}.reasoning_effort "${a}" is not one of its reasoning_efforts (${e.join(", ")})

${s}.model_provider names "${i}" but [model_providers.${i}] is not defined

${s}.base_url is required (on the model or its model_provider; api_base_url is also accepted)

${s}.api_backend is required — Grok Build defaults to chat_completions, which private inference does not speak; set api_backend = "${_}" on the model or its model_provider

${s}.api_backend = "${u}" is not supported — private inference speaks the OpenAI Responses API only; set api_backend = "${_}"

${s}.env_key names ${d.join(", ")} but none is set, and no auth_provider or [auth] auth_provider_command is configured

${r}.auth_provider names "${e}" but [auth_provider.${e}] is not defined

model_providers must be a table of [model_providers.<name>] tables

model_providers.${r} must be a table

auth_provider must be a table of [auth_provider.<name>] tables

auth_provider.${r} must be a table

auth_provider.${r}.command is required

auth must be a table

auth.auth_token_ttl is set but auth.auth_provider_command is not

private inference reads a Grok Build config.toml: one [model.<key>] table per model and [models] default = "<key>"

${t} is not valid TOML: ${(e instanceof Error?e.message:String(e)).split("\n",1)[0]}

${t} must contain a TOML document

models must be a [models] table, not a list — ${G}

no [model.<key>] tables — ${G}

model.${e} must be a table

models.default = "${a}" is not one of the [model.<key>] tables (${i.join(", ")})

models.default is required when more than one model is defined — name one of: ${i.join(", ")}

both a config file (${e.configFilePath}) and shorthand env vars (${r.join(", ")}) are set — ambiguous config; use one source

CURSOR_PRIVATE_INFERENCE_BASE_URL is not set

CURSOR_PRIVATE_INFERENCE_MODEL is not set

both CURSOR_PRIVATE_INFERENCE_API_KEY and CURSOR_PRIVATE_INFERENCE_API_KEY_HELPER are set — ambiguous credential config; set exactly one

CURSOR_PRIVATE_INFERENCE_TOKEN_LIMIT must be a positive integer

CURSOR_PRIVATE_INFERENCE_API_SHAPE is no longer supported — private inference speaks the OpenAI Responses API only

could not read ${e.configFilePath}

${n}: JSON private-inference config is not supported — the config is a Grok Build config.toml; point CURSOR_PRIVATE_INFERENCE_CONFIG at a .toml file

${n}: CURSOR_PRIVATE_INFERENCE_CONFIG must point at a .toml file

model "${e.id}": reasoning is declared without [[model.${e.id}.reasoning_efforts]] rungs and the gateway's /models advertises no tiers

model "${e.id}": reasoning_effort "${o}" is not one of the tiers the gateway advertises (${n.join(", ")})

model "${e.id}": the gateway advertises chat_completions for this model — private inference is Responses-only; the gateway must serve the Responses API

model "${e.id}": context_window is not configured and the gateway's /models advertises no context_length — set context_window (CURSOR_PRIVATE_INFERENCE_TOKEN_LIMIT for the shorthand)

model "${e.id}": long_context_threshold_tokens (${e.longContextThresholdTokens}) exceeds context_window (${r})

use strict

use strict

unsigned long long

long long

use strict

HTTP URIs must have a host.

URN can not be parsed

URN can not be parsed.

UUID is not valid.

use strict

use strict

ready state

binary type

sent close

received close

byte parser

use strict

use strict

use strict

Object expected.

Symbol.asyncDispose is not defined.

Symbol.dispose is not defined.

Object not disposable.

An error was suppressed during disposal.