@link-assistant/agent 0.20.2 → 0.22.0

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@link-assistant/agent",
-  "version": "0.20.2",
+  "version": "0.22.0",
   "description": "A minimal, public domain AI CLI agent compatible with OpenCode's JSON interface. Bun-only runtime.",
   "main": "src/index.js",
   "type": "module",

package/src/cli/argv.ts

@@ -105,7 +105,7 @@ export function getCompactionSafetyMarginFromProcessArgv(): string | null {
 /**
  * Extract --compaction-models argument directly from process.argv
  * The value is a links notation references sequence, e.g.:
- *   "(big-pickle nemotron-3-super-free minimax-m2.5-free gpt-5-nano qwen3.6-plus-free same)"
+ *   "(big-pickle minimax-m2.5-free nemotron-3-super-free gpt-5-nano same)"
  * @returns The compaction models argument from CLI or null if not found
  * @see https://github.com/link-assistant/agent/issues/232
  */

package/src/cli/continuous-mode.js

@@ -194,7 +194,8 @@ export async function runContinuousServerMode(
   systemMessage,
   appendSystemMessage,
   jsonStandard,
-  compactionModel
+  compactionModel,
+  temperature
 ) {
   // Check both CLI flag and environment variable for compact JSON mode
   const compactJson = argv['compact-json'] === true || config.compactJson;
@@ -290,6 +291,7 @@ export async function runContinuousServerMode(
             compactionModel,
             system: systemMessage,
             appendSystem: appendSystemMessage,
+            temperature,
           }),
         }
       ).catch((error) => {
@@ -446,7 +448,8 @@ export async function runContinuousDirectMode(
   systemMessage,
   appendSystemMessage,
   jsonStandard,
-  compactionModel
+  compactionModel,
+  temperature
 ) {
   // Check both CLI flag and environment variable for compact JSON mode
   const compactJson = argv['compact-json'] === true || config.compactJson;
@@ -523,6 +526,7 @@ export async function runContinuousDirectMode(
         compactionModel,
         system: systemMessage,
         appendSystem: appendSystemMessage,
+        temperature,
       }).catch((error) => {
         hasError = true;
         eventHandler.output({

package/src/cli/defaults.ts

@@ -6,7 +6,7 @@
  */
 
 /** Default model used when no `--model` CLI argument is provided. */
-export const DEFAULT_MODEL = 'opencode/qwen3.6-plus-free';
+export const DEFAULT_MODEL = 'opencode/nemotron-3-super-free';
 
 /** Default provider ID extracted from DEFAULT_MODEL. */
 export const DEFAULT_PROVIDER_ID = DEFAULT_MODEL.split('/')[0];
@@ -31,20 +31,21 @@ export const DEFAULT_COMPACTION_MODEL = 'opencode/gpt-5-nano';
  * The special value "same" means use the same model as `--model`.
  *
  * Parsed as links notation references sequence (single anonymous link):
- *   "(big-pickle nemotron-3-super-free minimax-m2.5-free gpt-5-nano qwen3.6-plus-free same)"
+ *   "(big-pickle minimax-m2.5-free nemotron-3-super-free gpt-5-nano same)"
  *
  * Context limits (approximate):
  *   big-pickle:            ~200K
- *   nemotron-3-super-free: ~262K
  *   minimax-m2.5-free:     ~200K
+ *   nemotron-3-super-free: ~262K (default model)
  *   gpt-5-nano:            ~400K
- *   qwen3.6-plus-free:     ~1M
  *   same:                  (base model's context)
  *
+ * Note: qwen3.6-plus-free was removed — free promotion ended April 2026.
+ * @see https://github.com/link-assistant/agent/issues/242
  * @see https://github.com/link-assistant/agent/issues/232
  */
 export const DEFAULT_COMPACTION_MODELS =
-  '(big-pickle nemotron-3-super-free minimax-m2.5-free gpt-5-nano qwen3.6-plus-free same)';
+  '(big-pickle minimax-m2.5-free nemotron-3-super-free gpt-5-nano same)';
 
 /**
  * Default compaction safety margin as a percentage of usable context window.

package/src/cli/run-options.js

@@ -168,5 +168,10 @@ export function buildRunOptions(yargs) {
       description:
         'Safety margin (%) of usable context window before triggering compaction. Only applies when the compaction model has equal or smaller context than the base model. Default: 15.',
       default: DEFAULT_COMPACTION_SAFETY_MARGIN_PERCENT,
+    })
+    .option('temperature', {
+      type: 'number',
+      description:
+        'Override the temperature for model completions. When not set, the default per-model temperature is used.',
     });
 }

package/src/index.js

@@ -313,7 +313,8 @@ async function runAgentMode(argv, request) {
           systemMessage,
           appendSystemMessage,
           jsonStandard,
-          compactionModel
+          compactionModel,
+          argv.temperature
         );
       } else {
         // DIRECT MODE: Run everything in single process
@@ -325,7 +326,8 @@ async function runAgentMode(argv, request) {
           systemMessage,
           appendSystemMessage,
           jsonStandard,
-          compactionModel
+          compactionModel,
+          argv.temperature
         );
       }
     },
@@ -399,7 +401,8 @@ async function runContinuousAgentMode(argv) {
           systemMessage,
           appendSystemMessage,
           jsonStandard,
-          compactionModel
+          compactionModel,
+          argv.temperature
         );
       } else {
         // DIRECT MODE: Run everything in single process
@@ -410,7 +413,8 @@ async function runContinuousAgentMode(argv) {
           systemMessage,
           appendSystemMessage,
           jsonStandard,
-          compactionModel
+          compactionModel,
+          argv.temperature
         );
       }
     },
@@ -433,7 +437,8 @@ async function runServerMode(
   systemMessage,
   appendSystemMessage,
   jsonStandard,
-  compactionModel
+  compactionModel,
+  temperature
 ) {
   const compactJson = argv['compact-json'] === true;
 
@@ -502,6 +507,7 @@ async function runServerMode(
           compactionModel,
           system: systemMessage,
           appendSystem: appendSystemMessage,
+          temperature,
         }),
       }
     ).catch((error) => {
@@ -534,7 +540,8 @@ async function runDirectMode(
   systemMessage,
   appendSystemMessage,
   jsonStandard,
-  compactionModel
+  compactionModel,
+  temperature
 ) {
   const compactJson = argv['compact-json'] === true;
 
@@ -587,6 +594,7 @@ async function runDirectMode(
       compactionModel,
       system: systemMessage,
       appendSystem: appendSystemMessage,
+      temperature,
     }).catch((error) => {
       hasError = true;
       eventHandler.output({

package/src/provider/provider.ts

@@ -1749,10 +1749,9 @@ export namespace Provider {
     }
     if (providerID === 'opencode' || providerID === 'local') {
       priority = [
-        'qwen3.6-plus-free',
+        'nemotron-3-super-free',
         'minimax-m2.5-free',
         'gpt-5-nano',
-        'nemotron-3-super-free',
         'big-pickle',
       ];
     }
@@ -1781,9 +1780,8 @@ export namespace Provider {
   }
 
   const priority = [
-    'qwen3.6-plus-free',
-    'glm-5-free',
     'nemotron-3-super-free',
+    'glm-5-free',
     'minimax-m2.5-free',
     'gpt-5-nano',
     'big-pickle',
@@ -1866,7 +1864,7 @@ export namespace Provider {
    * 1. If model is uniquely available in one provider, use that provider
    * 2. If model is available in multiple providers, prioritize based on free model availability:
    *    - kilo: glm-5-free, glm-4.5-air-free, minimax-m2.5-free, giga-potato-free, deepseek-r1-free (unique to Kilo)
-   *    - opencode: big-pickle, gpt-5-nano, qwen3.6-plus-free, nemotron-3-super-free (unique to OpenCode)
+   *    - opencode: big-pickle, gpt-5-nano, nemotron-3-super-free (unique to OpenCode)
    * 3. For shared models, prefer OpenCode first, then fall back to Kilo on rate limit
    *
    * @param modelID - Short model name without provider prefix

package/src/session/message-v2.ts

@@ -411,6 +411,7 @@ export namespace MessageV2 {
       .optional(),
     system: z.string().optional(),
     appendSystem: z.string().optional(),
+    temperature: z.number().optional(),
     tools: z.record(z.string(), z.boolean()).optional(),
   }).meta({
     ref: 'UserMessage',

package/src/session/prompt.ts

@@ -110,6 +110,7 @@ export namespace SessionPrompt {
     noReply: z.boolean().optional(),
     system: z.string().optional(),
     appendSystem: z.string().optional(),
+    temperature: z.number().optional(),
     tools: z.record(z.string(), z.boolean()).optional(),
     parts: z.array(
       z.discriminatedUnion('type', [
@@ -734,10 +735,12 @@ export namespace SessionPrompt {
       });
       const params = {
         temperature:
-          (model.info?.temperature ?? false)
-            ? (agent.temperature ??
-              ProviderTransform.temperature(model.providerID, model.modelID))
-            : undefined,
+          lastUser.temperature != null
+            ? lastUser.temperature
+            : (model.info?.temperature ?? false)
+              ? (agent.temperature ??
+                ProviderTransform.temperature(model.providerID, model.modelID))
+              : undefined,
         topP:
           agent.topP ?? ProviderTransform.topP(model.providerID, model.modelID),
         options: {
@@ -1189,6 +1192,7 @@ export namespace SessionPrompt {
       tools: input.tools,
       system: input.system,
       appendSystem: input.appendSystem,
+      temperature: input.temperature,
       agent: agent.name,
       model: await resolveModel({
         model: input.model,