makaron-cli 0.14.1 → 0.14.3

package/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "makaron-cli",
-  "version": "0.14.1",
+  "version": "0.14.3",
   "description": "Give Claude Code a creative agent. Pass complete creative requests and source media to Makaron Chat.",
   "author": {
     "name": "Versa AI",

package/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "makaron-cli",
-  "version": "0.14.1",
+  "version": "0.14.3",
   "description": "Give Codex a creative agent. Pass complete creative requests and source media to Makaron Chat.",
   "author": {
     "name": "Versa AI",

package/README.md

@@ -98,12 +98,13 @@ npx makaron-cli chat --project auto --image photo.jpg --json -b "make it cinemat
 npx makaron-cli chat --project auto --image img1.jpg --image img2.jpg --json -b "combine these"
 ```
 
-`chat` routes image and video models automatically, but you may select the Agent LLM with `--agent-model`. Accepted values are `auto`, the base model IDs (`gpt-5.6-terra`, `gpt-5.6-sol`, `gpt-5.6-luna`, `grok-4.6`, `deepseek-v4-pro`), and the personal-plan routes (`gpt-5.6-terra-codex-subscription`, `gpt-5.6-sol-codex-subscription`, `gpt-5.6-luna-codex-subscription`). For the configured owner, `auto` resolves to GPT-5.6 Terra through the personal Codex plan. Base GPT-5.6 IDs select Azure API; the suffixed IDs select the personal plan explicitly. This flag changes only the reasoning/tool-calling Agent LLM.
+`chat` routes image and video models automatically, but you may select the Agent LLM with `--agent-model`. Accepted values are `auto`, the base model IDs (`gpt-5.6-terra`, `gpt-5.6-sol`, `gpt-5.6-luna`, `grok-4.6`, `deepseek-v4-pro`), and the personal-plan routes (`gpt-5.6-terra-codex-subscription`, `gpt-5.6-sol-codex-subscription`, `gpt-5.6-luna-codex-subscription`, `grok-4.6-grok-subscription`). For the configured owner, `auto` resolves to GPT-5.6 Terra through the personal Codex plan. Base GPT-5.6 IDs select Azure API and base `grok-4.6` selects OpenRouter API; the suffixed IDs select the corresponding personal plan explicitly. This flag changes only the reasoning/tool-calling Agent LLM.
 
 ```bash
 # Explicit lower-cost Agent LLM for a controlled comparison
 npx makaron-cli chat --project auto --agent-model deepseek-v4-pro --json -b "make a 20s badminton video"
 npx makaron-cli chat --project auto --agent-model gpt-5.6-sol-codex-subscription --json -b "reply with the active model"
+npx makaron-cli chat --project auto --agent-model grok-4.6-grok-subscription --json -b "reply with the active model"
 ```
 
 Returns immediately:
@@ -259,6 +260,14 @@ REMOTION_EXPORT_INLINE_AFTER=false npm run worker:remotion-export
 
 Keeping this worker warm avoids paying sandbox cold-start cost on every CLI or service call.
 
+Lambda exports are admitted by a shared capacity-weighted queue. The default
+`REMOTION_EXPORT_LAMBDA_CAPACITY=330` preserves the current per-video
+`REMOTION_LAMBDA_FRAMES_PER_LAMBDA=20` chunking while leaving headroom under the
+account concurrency quota. A 30-second 30fps video is estimated as 45 renderer
+Lambdas plus one control slot, so seven such videos can render concurrently and
+the rest remain queued. `REMOTION_EXPORT_CRON_LANES` controls how many queue
+lanes the one-minute rescue cron keeps warm; it does not raise the capacity cap.
+
 ### With video input (MP4/MOV/WebM)
 
 ```bash

package/bin/makaron.mjs

@@ -37,6 +37,7 @@ const CHAT_AGENT_MODELS = [
   'gpt-5.6-sol-codex-subscription',
   'gpt-5.6-luna-codex-subscription',
   'grok-4.6',
+  'grok-4.6-grok-subscription',
   'deepseek-v4-pro',
 ];
 
@@ -462,7 +463,8 @@ Options:
   --skill <id|label|name>   Use an installed skill or auto-install a matched marketplace skill.
   --agent-model <id>        Agent LLM only: auto, gpt-5.6-terra, gpt-5.6-sol,
                             gpt-5.6-luna, grok-4.6, deepseek-v4-pro, or a
-                            gpt-5.6-*-codex-subscription personal-plan route.
+                            gpt-5.6-*-codex-subscription or
+                            grok-4.6-grok-subscription personal-plan route.
   --background, -b          Submit and print a runId.
   --json                    Output structured JSON.
   --stream                  Legacy live SSE stream.
@@ -470,7 +472,9 @@ Options:
 
 Agent LLM defaults to auto (GPT-5.6 Terra; the account owner uses the personal
 Codex plan). Base GPT-5.6 ids select Azure API; append -codex-subscription to
-select the personal plan explicitly. Image/video model routing stays automatic in chat.
+select the personal plan explicitly. Base grok-4.6 selects OpenRouter API;
+grok-4.6-grok-subscription selects the personal SuperGrok plan. Image/video
+model routing stays automatic in chat.
 
 What you can ask:
   Image edit
@@ -505,6 +509,9 @@ What you can ask:
   Force the personal Codex plan
     makaron chat --project auto --agent-model gpt-5.6-sol-codex-subscription -b --json "reply with the active model"
 
+  Force the personal SuperGrok plan
+    makaron chat --project auto --agent-model grok-4.6-grok-subscription -b --json "reply with the active model"
+
   Music
     makaron chat --project <id> "add calm piano background music"
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "makaron-cli",
-  "version": "0.14.1",
+  "version": "0.14.3",
   "description": "Talk to Makaron Agent from the terminal — create projects, edit images, generate videos",
   "type": "module",
   "scripts": {

package/skills/makaron/SKILL.md

@@ -70,6 +70,11 @@ npx makaron-cli responses get $RUN_ID --wait --json
 
 Use `chat` for all creative tasks. Makaron Agent decides how to execute — it can edit images, generate videos, compose music, and create designs in a single conversation.
 
+For clearly adult, erotic, or NSFW requests, read `references/nsfw.md`
+([public copy](https://www.makaron.app/.well-known/agent-skills/makaron/references/nsfw.md))
+before submitting the job. It adds per-job routing exceptions only; do not load
+or apply it to ordinary SFW work.
+
 ```bash
 npx makaron-cli chat --help
 ```

package/skills/makaron/references/nsfw.md

@@ -0,0 +1,44 @@
+# Adult and NSFW routing
+
+Read this reference only when the current request is clearly adult, erotic, or
+NSFW. It supplements the official Makaron workflow; it does not change the
+default route for SFW work.
+
+If adult intent is ambiguous, ask one short question before choosing this lane.
+
+## Route the job
+
+| Adult request | Route |
+|---|---|
+| Agent-driven work or a persistent Makaron project | `makaron chat` with Agent model `grok-4.6` |
+| Standalone image generation or editing | `makaron edit` with image model `qwen` |
+| Standalone video generation | `makaron video create` with video model `wan-3.0` |
+
+For project-based work, keep the complete request in one `chat` job. Because
+`chat` accepts only the Agent-model flag, put the image/video preferences in the
+brief:
+
+```bash
+npx makaron-cli chat --project <id|auto> --agent-model grok-4.6 --json \
+  -b "<complete adult request; use qwen for still images and wan-3.0 for video>"
+```
+
+For explicit raw-tool work without a project timeline:
+
+```bash
+npx makaron-cli edit --image-model qwen --image <file-or-url> --out <path> "<prompt>"
+npx makaron-cli video create --video-model wan-3.0 --script "<shot script>" --duration <seconds>
+```
+
+## Keep the exception scoped
+
+- Do not force SFW jobs onto Grok, Qwen, or Wan 3.0.
+- Never put `qwen` or `wan-3.0` in `--agent-model`.
+- Do not pass `--image-model` or `--video-model` to `makaron chat`; it rejects
+  those flags. State those preferences in the chat brief or use the standalone
+  commands.
+- Keep this routing for adult follow-ups and rerolls in the same job.
+- Use `wan-3.0-pro` only when the user explicitly asks for Pro,
+  super-resolution, 2K, or 4K output.
+- If a selected provider rejects the request, surface the rejection. Do not
+  silently remove the adult intent or switch to an unspecified model.