@remixmate/cli 0.1.0 → 0.1.1

package/skills/gen-video/SKILL.md

@@ -17,11 +17,13 @@ triggers:
 
 # AI Video Generation Skill
 
-Wraps ab-api's `POST /model/genVideo` (the same endpoint the web "Lingchuang AI Video" tool uses), authenticated with the **Tianyan privateToken**, routed through LiteLLM to **Seedance** or **Veo**.
+Wraps ab-api's `POST /model/genVideo` (the same endpoint the web "Lingchuang AI Video" tool uses), authenticated with the **Tianyan privateToken**, routed through LiteLLM to **Seedance** or **Veo**. Generation is async — the handler submits the task and polls `/model/getVideoStatus` until completion.
+
+> This skill was migrated from a Python script to an ab-skill CLI HTTP handler (`entry.type: http`). The agent invocation is unchanged (same tool name `gen_video`, same params as in `skill.json`); local repro goes through `ab-skill gen-video ...`.
 
 ## Models
 
-Aligned with `gen_video.py`'s `GEN_VIDEO_MODEL_PRESETS` and the frontend `AI_VIDEO_MODELS`:
+Aligned with the handler's `MODEL_ALIASES` and the frontend `AI_VIDEO_MODELS`:
 
 | LiteLLM `model` | Display name | Provider | Duration | Notes |
 |-----------------|--------------|----------|----------|-------|
@@ -49,29 +51,28 @@ Aligned with `gen_video.py`'s `GEN_VIDEO_MODEL_PRESETS` and the frontend `AI_VID
 
 ## Auth & environment
 
-No skill-local env file — the executing process inherits the system environment. Examples below say `python`; on macOS you may need `python3`.
+No skill-local env file — the executing process inherits the system environment.
 
 - **Enterprise OpenClaw**: auth is already injected, **no need** for `PRIV_TOKEN` / `--priv-token`.
-- **Other environments**: configure the token. Without a token, non-interactive runs fail; interactive ones may prompt.
+- **Other environments**: configure the token. Without a token, non-interactive runs fail.
 
 | Env var | Description | Default |
 |---------|-------------|---------|
 | `PRIV_TOKEN` | Tianyan token; `--priv-token` overrides | none |
 | `MM_VIDEO_MODEL` | Default model id | `doubao-seedance-1-5-pro-251215` |
-| `MM_API_BASE_URL` | API root | see `gen_video.py` default |
+| `MM_API_BASE_URL` | API root; `--api-base-url` overrides | `http://localhost:2999/api` |
 | `AGENT_NAME` | Optional `x-invoke-agent` header | none |
 
 ## Operations
 
-1. **Script path**: read the system-injected `Base directory for this skill: <path>` as `<SkillDir>` and always use `<SkillDir>/scripts/gen_video.py`. Do not hardcode paths.
-2. **Prompt**: include camera language ("aerial shot", "slow motion", "push-in") and dynamic descriptions for better results. Chinese or English both work.
-3. By default only the URL is printed; pass `--download` (and optional `--output`) to save locally.
-4. **Async job**: video generation is async; the script polls until completion (typically 1–3 minutes).
+1. **Prompt**: include camera language ("aerial shot", "slow motion", "push-in") and dynamic descriptions for better results. Chinese or English both work. For pure image-to-video, an empty prompt with first/last frame is allowed.
+2. **Async job**: video generation is async; the handler polls until completion (typically 1–3 minutes) and emits `__progress__` lines.
+3. **Surface results**: stdout prints the video URL on its own line; show it directly to the user (no download needed — the URL is cloud-persisted).
 
 ### Seedance text-to-video (default model)
 
 ```bash
-python3 <SkillDir>/scripts/gen_video.py \
+ab-skill gen-video \
   --prompt "<video description>" \
   --duration 5 \
   --ratio "16:9"
@@ -79,8 +80,10 @@ python3 <SkillDir>/scripts/gen_video.py \
 
 ### Seedance first/last frame + audio
 
+Frame images accept local file paths, HTTPS URLs, or data URIs (local files are base64-encoded into a data URI).
+
 ```bash
-python3 <SkillDir>/scripts/gen_video.py \
+ab-skill gen-video \
   --prompt "<transition description>" \
   --first-frame ./start.png \
   --last-frame ./end.png \
@@ -91,7 +94,7 @@ python3 <SkillDir>/scripts/gen_video.py \
 ### Veo 3.1 high-resolution
 
 ```bash
-python3 <SkillDir>/scripts/gen_video.py \
+ab-skill gen-video \
   --model veo \
   --prompt "<video description>" \
   --duration 8 \
@@ -102,7 +105,7 @@ python3 <SkillDir>/scripts/gen_video.py \
 ### Veo 3.1 Fast for rapid iteration
 
 ```bash
-python3 <SkillDir>/scripts/gen_video.py \
+ab-skill gen-video \
   --model veo-fast \
   --prompt "<video description>" \
   --duration 6
@@ -110,8 +113,10 @@ python3 <SkillDir>/scripts/gen_video.py \
 
 ### Veo with reference images
 
+Pass `--reference` multiple times (Veo only, up to 3).
+
 ```bash
-python3 <SkillDir>/scripts/gen_video.py \
+ab-skill gen-video \
   --model veo \
   --prompt "<video description>" \
   --reference ./ref1.png \
@@ -119,19 +124,6 @@ python3 <SkillDir>/scripts/gen_video.py \
   --duration 8
 ```
 
-### Download to disk
-
-```bash
-python3 <SkillDir>/scripts/gen_video.py \
-  --prompt "<video description>" \
-  --download \
-  --output "output_video.mp4"
-```
-
-5. **Surface results**: without `--download`, print the video URL; with `--download`, mention the local path.
-
-For the full flag list: `python3 <SkillDir>/scripts/gen_video.py --help`.
-
 ## Common CLI flags
 
 | Flag | Description | Default |
@@ -149,11 +141,8 @@ For the full flag list: `python3 <SkillDir>/scripts/gen_video.py --help`.
 | `--negative-prompt` | Negative prompt (mainly Veo) | none |
 | `--seed` | Random seed | none |
 | `--person-generation` | Person policy: `allow_all` / `dont_allow` (Veo) | none |
-| `--download` | Download the video to disk | off |
-| `-o` / `--output` | Download path (only effective with `--download`) | `generated_video.mp4` |
-| `--poll-interval` | Polling interval (seconds) | `10` |
-| `--max-wait` | Max wait time (seconds) | `300` |
-| `--private-token` | Override token | see above |
+| `--api-base-url` | Override API root | see above |
+| `--priv-token` | Override token | see above |
 
 ## Prompt tips
 
@@ -166,5 +155,5 @@ For the full flag list: `python3 <SkillDir>/scripts/gen_video.py --help`.
 - **401** / **token missing** (non-OpenClaw): set `PRIV_TOKEN`.
 - **Business `code != 0`**: read `msg` on stderr.
 - **429**: rate-limited; retry later.
-- **Generation timeout**: increase `--max-wait` or retry later.
+- **Generation timeout**: the handler caps polling at ~5 minutes; retry later if it times out.
 - **Network**: verify connectivity and `MM_API_BASE_URL`.

package/skills/gen-video/skill.json

@@ -5,7 +5,7 @@
   "title": "AI Video Generation",
   "description": "AI video generation: produce a short video clip from a text prompt. Supports Seedance and Veo models, plus first/last frame and reference images.",
   "envVars": ["PRIV_TOKEN", "MM_API_BASE_URL", "AGENT_NAME", "MM_VIDEO_MODEL"],
-  "scriptPath": "scripts/gen_video.py",
+  "entry": { "type": "http", "handler": "gen-video" },
   "parameters": {
     "type": "object",
     "properties": {
@@ -17,7 +17,8 @@
       "first_frame": { "type": "string", "description": "First-frame image path or URL" },
       "last_frame": { "type": "string", "description": "Last-frame image path or URL" },
       "reference": { "type": "string", "description": "Reference image (Veo only)" },
-      "generate_audio": { "type": "boolean", "description": "Generate native audio" }
+      "generate_audio": { "type": "boolean", "description": "Generate native audio" },
+      "json_output": { "type": "boolean", "description": "Emit a JSON result ({ url }) instead of human-readable output" }
     },
     "required": ["prompt"]
   }

package/skills/gen-video/version.json

@@ -2,6 +2,6 @@
   "skillName": "gen-video",
   "repoName": "agent-skill-media-maker",
   "skillId": "339",
-  "version": "V5",
+  "version": "V7",
   "skillDescription": "AI 生视频技能，根据文字描述生成素材视频（调用 ab-api /model/genVideo，支持 Seedance 与 Veo）。\n\n当用户提到以下任何需求时，立即使用本 skill：\n- AI 生视频、文生视频、文字生成视频、生成一段视频、AI 制作视频\n- 使用 doubao / 豆包 / seedance、Veo、Google 等生成视频\n- 用户提供视频提示词并希望生成视频\n- 图生视频、首帧生成视频、参考图生成视频\n\n即使用户没有明确说「使用 AI」，只要他们想要根据描述生成视频，也要使用本 skill。"
 }

package/skills/gen-voice/SKILL.md

@@ -29,7 +29,7 @@ There is no skill-local env file — the executing process inherits the system e
 | Env var | Description | Default |
 |---------|-------------|---------|
 | `PRIV_TOKEN` | Tianyan token; `--priv-token` overrides | none |
-| `MM_API_BASE_URL` | API root | see `gen_voice.py` default |
+| `MM_API_BASE_URL` | API root; `--api-base-url` overrides | `http://localhost:2999/api` |
 | `AGENT_NAME` | Optional `x-invoke-agent` header | none |
 
 ## Operations

package/skills/gen-voice/version.json

@@ -2,6 +2,6 @@
   "skillName": "gen-voice",
   "repoName": "agent-skill-media-maker",
   "skillId": "338",
-  "version": "V5",
+  "version": "V6",
   "skillDescription": "语音合成（TTS）技能，将文字转为语音音频（调用 ab-api /tool/minimaxT2aV2，基于 Minimax TTS）。\n\n当用户提到以下任何需求时，立即使用本 skill：\n- AI 配音、语音合成、文字转语音、TTS、文本转音频、生成语音\n- 使用 minimax 语音合成\n- 用户想要将一段文字朗读出来、生成音频\n\n即使用户没有明确说「使用 AI」，只要他们想要将文字转为语音，也要使用本 skill。"
 }

package/skills/prepare-video-assets/SKILL.md

@@ -79,7 +79,7 @@ python3 <SkillDir>/scripts/prepare_video_assets.py \
   --template-id html-slide
 ```
 
-`--template-id` automatically calls template-bind internally; or pass `--binding my-video.binding.json` if you have a pre-generated binding.
+`--template-id` automatically calls template-registry internally; or pass `--binding my-video.binding.json` if you have a pre-generated binding.
 
 ## Image-URL handling (agent rules)
 
@@ -189,4 +189,4 @@ The Asset Resolver handles each `AssetRef` in this order:
 ## See also
 
 - **`render_video`** — the Phase 3 skill that consumes this skill's `job_id` and produces the final video.
-- **`template-bind`** — Lists available templates. Called internally by this skill when `--template-id` is provided.
+- **`template-registry`** — Lists available templates. Called internally by this skill when `--template-id` is provided.

package/skills/prepare-video-assets/skill.json

@@ -11,7 +11,7 @@
     "properties": {
       "dsl_json": { "type": "string", "description": "DSL JSON as an inline string (preferred — no disk write needed; required for multi-user concurrent flows). Pass the full DSL JSON, or — when a gen_script skeleton is cached in the session — a minimal JSON with only scenes[].audio.narration.text overrides (the agent layer merges narration onto the cached skeleton)." },
       "dsl": { "type": "string", "description": "DSL file path (legacy fallback; prefer dsl_json inline)." },
-      "template_id": { "type": "string", "description": "Template id. When provided, the script invokes template-bind internally; no separate binding step is needed." },
+      "template_id": { "type": "string", "description": "Template id. When provided, the script invokes template-registry internally; no separate binding step is needed." },
       "binding_json": { "type": "string", "description": "TemplateBinding JSON as an inline string (alternative to template_id; no file needed)." },
       "binding": { "type": "string", "description": "TemplateBinding file path (optional; auto-generated when template_id is provided)." },
       "save_job": { "type": "boolean", "description": "Persist the RenderPlan to the database (default true). On success the stdout contains '📦 render job jobId: N'. Auto-degrades to false when PRIV_TOKEN is missing, falling back to file mode. Pass false to opt out explicitly." },

package/skills/render-video/scripts/render_video.py

@@ -38,17 +38,17 @@ from datetime import datetime, timezone
 from typing import Optional
 
 # ── Cross-skill dependency ────────────────────────────────────────────────
-# Reach into `skills/template-bind/` — that directory is the de-facto home
+# Reach into `skills/template-registry/` — that directory is the de-facto home
 # for shared Python code in ab-skill (timeline-compilation, registry loader,
-# render-job HTTP client, etc.). Naming it `_SHARED_LIB_DIR` rather than
-# `_TEMPLATE_BIND_DIR` reflects its actual role: it hosts much more than
-# template-binding now (see AGENTS.md "skills/template-bind/scripts/" note).
+# render-job HTTP client, etc.). Naming it `_SHARED_LIB_DIR` reflects its
+# actual role: it hosts much more than template lookup (see AGENTS.md
+# "skills/template-registry/scripts/" note).
 #
 # If signatures of the functions below change, also update:
-#   - skills/template-bind/video_dsl/runtime/timeline_compiler.py (the source)
+#   - skills/template-registry/video_dsl/runtime/timeline_compiler.py (the source)
 #   - any other caller discoverable via `grep -r "split_subtitle\|segment_narration"`
 # ───────────────────────────────────────────────────────────────────────────
-_SHARED_LIB_DIR = os.path.join(os.path.dirname(__file__), "..", "..", "template-bind")
+_SHARED_LIB_DIR = os.path.join(os.path.dirname(__file__), "..", "..", "template-registry")
 # Single sys.path setup for both cross-skill import surfaces of the shared lib:
 #   - <shared>            for `video_dsl.runtime.*`
 #   - <shared>/scripts    for `registry_loader`, `match_template`,
@@ -1779,7 +1779,7 @@ def auto_bind_template(dsl: dict, template_id: str) -> dict:
     """Auto-generate TemplateBinding from DSL + template-id by importing
     match_template.py's binding logic. Eliminates the need for a separate
     match_template.py invocation step."""
-    # template-bind/scripts is on sys.path via the module-top setup.
+    # template-registry/scripts is on sys.path via the module-top setup.
     try:
         import match_template
     except ImportError as e:
@@ -1885,7 +1885,7 @@ Examples:
         if not private_token:
             LogPrint("❌ --job-id mode requires PRIV_TOKEN or --priv-token", file=sys.stderr)
             sys.exit(1)
-        # render_job_client lives in skills/template-bind/scripts/ (the shared
+        # render_job_client lives in skills/template-registry/scripts/ (the shared
         # location for cross-skill helpers); it is already on sys.path via the
         # module-top setup.
         from render_job_client import get_plan as rjc_get_plan

package/skills/{template-bind → template-registry}/README.md

@@ -1,4 +1,4 @@
-# template-bind
+# template-registry
 
 CLI 暴露面是「列出可用模板」(`--list-templates`)。把 Video DSL 场景绑到具体模板 / 生成 TemplateBinding 的 Python 逻辑由本 skill 提供为**库**(`scripts/match_template.py`),被 `render-video` / `prepare-video-assets` 通过 `sys.path` 内联 import,**没有独立 CLI**。
 
@@ -20,7 +20,7 @@ ab-api 响应包装格式 `{code, msg, data}` 由 `registry_loader._fetch_http`
 
 ## 新增模板工作流
 
-新增模板**不需要改 ab-skill / template-bind 代码**:
+新增模板**不需要改 ab-skill / template-registry 代码**:
 
 1. 在 template-library 仓库定义新模板(`template.json` + 组件代码)。
 2. template-library CI 校验 schema + 契约。
@@ -50,13 +50,13 @@ python3 <SkillDir>/scripts/check_contracts.py
 | 模块 | 谁在用 | 作用 |
 |---|---|---|
 | `template_paths.py` | registry_loader / template_binder / gen-script / check_contracts | 统一计算 monorepo / template-library 路径,代替 4 个 caller 各自数 `..` |
-| `registry_loader.py` | template-bind handler / match_template / render-video / gen-script / dsl_validator | 加载 registry(HTTP + cache + file) |
+| `registry_loader.py` | template-registry / match_template / render-video / gen-script / dsl_validator | 加载 registry(HTTP + cache + file) |
 | `match_template.py` | render-video(`auto_bind_template`) | DSL → TemplateBinding 算法 |
 | `render_job_client.py` | render-video / export-jianying | ab-api renderJob HTTP 客户端 |
 | `check_contracts.py` | 维护脚本 | template-library ↔ remotion-renderer 契约校验 |
 | `sync_registry.py` | 维护脚本 | 把 registry 摘要落盘供 LLM prompt 使用 |
 
-sibling skill 通过 `sys.path.insert(0, "<…>/template-bind/scripts")` 然后直接 `import` 模块名,模式见 `render_video.py` 顶部的"Single sys.path setup"段。
+sibling skill 通过 `sys.path.insert(0, "<…>/template-registry/scripts")` 然后直接 `import` 模块名,模式见 `render_video.py` 顶部的"Single sys.path setup"段。
 
 ## 详细使用说明
 

package/skills/{template-bind → template-registry}/SKILL.md

@@ -1,17 +1,14 @@
 ---
-name: template-bind
+name: template-registry
 description: |
-  Video-template registry and binding skill. Stores every video-template definition and produces a TemplateBinding for a given template id.
+  Video-template registry skill. Stores every video-template definition and lists the available templates (templateId / name / aspect ratio / style tags).
 
   Use this skill as soon as the user mentions any of these intents:
   - View available templates / list every template
-  - Bind a template to a DSL (a template-id must be supplied)
 
-  Also use this skill whenever the agent has a Video DSL and needs to produce a TemplateBinding.
+  Note: DSL→TemplateBinding is no longer a separate exposed step — once prepare_video_assets receives a template_id it builds the binding internally.
 triggers:
   - View available templates / list every template
-  - Bind a template to a DSL (a template-id must be supplied)
-  - Agent has a Video DSL and needs a TemplateBinding
 ---
 
 # Video Template Registry & Binding Skill
@@ -28,7 +25,7 @@ Stores every video-template definition, loads a template by **template-id**, and
 
 Template metadata lives in the standalone **template-library** monorepo and ships via the `@ab-templates/metadata` package.
 
-template-bind reads `template-library/packages/metadata/registry.json` to load every template definition (with the full slotMapping, compositions, etc.).
+template-registry reads `template-library/packages/metadata/registry.json` to load every template definition (with the full slotMapping, compositions, etc.).
 
 ```
 template-library/packages/
@@ -76,11 +73,11 @@ Resolution order (see `scripts/registry_loader.py` for the canonical implementat
 
 ## Steps
 
-> This skill is now a built-in handler of the ab-skill CLI (`entry.type: builtin`). The agent tool name `template_bind` is still the only entry; locally reproduce via `ab-skill template-bind --list-templates`.
+> This skill is a Python skill of the ab-skill CLI (`entry.type: python` → `scripts/list_templates.py`). The agent tool name `template_registry` is the only entry; locally reproduce via `ab-skill template-registry --list-templates`. The list command delegates to `scripts/registry_loader.py` — the same loader (with caching + stable/beta gating) that `render-video` and `gen-script` import in-process, so there is a single registry-reading implementation.
 >
-> The Python binding logic that maps DSL → TemplateBinding lives in `scripts/match_template.py` but is **no longer exposed as a CLI** — it is only consumed as a Python library by `render-video`'s `render_video.py` via `import match_template`. The `template-bind` skill itself no longer has a Python entry script.
+> The Python binding logic that maps DSL → TemplateBinding lives in `scripts/match_template.py` but is **not exposed as a CLI** — it is only consumed as a Python library by `render-video`'s `render_video.py` via `import match_template`.
 
-1. **List available templates**: run `ab-skill template-bind --list-templates` to view the templates in the registry along with their supported aspect ratios / style tags, and decide which `templateId` to pick.
+1. **List available templates**: run `ab-skill template-registry --list-templates` to view the templates in the registry along with their supported aspect ratios / style tags, and decide which `templateId` to pick.
 2. **Write the DSL**: when generating the Video DSL, put the chosen `templateId` into `meta.templateId` (the canonical location). Use `meta.templateVariant` / `renderHints.templateVariant` to explicitly select a variant. The legacy `renderHints.templatePreference[0]` is still tolerated by `match_template.py` and `dsl_validator._pick_template_id` during transition, but new authors should write `meta.templateId`.
 3. **Produce the TemplateBinding**: there is no standalone CLI for DSL → TemplateBinding; `prepare_video_assets` calls `match_template.build_binding(template, dsl)` inline during the asset-resolution pipeline and embeds the binding into the RenderPlan it hands to the renderer — no separate `*.binding.json` file is written.
 
@@ -89,7 +86,7 @@ Resolution order (see `scripts/registry_loader.py` for the canonical implementat
 ### List available templates
 
 ```bash
-ab-skill template-bind --list-templates
+ab-skill template-registry --list-templates
 ```
 
 ### Sync the registry cache (optional, used for offline / LLM prompt)
@@ -123,7 +120,12 @@ print(json.dumps(binding, ensure_ascii=False, indent=2))
 
 | Flag | Description | Default |
 |------|-------------|---------|
-| `--list-templates` | List every available template and exit (currently the only CLI verb). | — |
+| `--list-templates` | List every available template (currently the only CLI verb). | — |
+| `--filter-tag` | Keep only templates whose `styleTags` match this substring (case-insensitive). | none |
+| `--filter-aspect` | Keep only templates declaring this aspect ratio (e.g. `9:16`). | none |
+| `--filter-language` | Keep only templates whose `contentLanguage` includes this code (`zh`/`en`); language-agnostic templates always show. | none |
+| `--include-beta` | Also show `status: beta` templates (same effect as `ENABLE_BETA_TEMPLATES=1`). | off |
+| `--json-output` | Emit `{ "templates": [...] }` instead of the table. | off |
 
 ## Props extraction rules
 

package/skills/template-registry/scripts/list_templates.py

@@ -0,0 +1,129 @@
+#!/usr/bin/env python3
+"""
+template-registry CLI — list available video templates.
+
+Single source of truth: registry loading, caching and stable/beta status
+gating all live in ``registry_loader.py`` (the same module ``render-video``
+and ``gen-script`` import in-process). This entry point only adds the CLI
+surface — flag parsing, tag/aspect/language filtering, and table/JSON
+rendering — that used to be re-implemented in TypeScript
+(``src/handlers/template-bind.ts``). Deleting that handler removes the
+TS↔Python duplication; this is now the only place the list verb is realized.
+
+The DSL → TemplateBinding logic stays in ``match_template.py`` and is invoked
+in-process by ``render-video``; it is intentionally NOT exposed as a CLI verb.
+
+Flags:
+  --list-templates                 (accepted; listing is the only verb)
+  --filter-tag <substr>            keep templates whose styleTags match
+  --filter-aspect <ratio>          keep templates supporting this aspect ratio
+  --filter-language <zh|en|...>    keep templates tagged with this content lang
+                                   (templates with no contentLanguage are kept)
+  --include-beta                   also show beta templates (else stable-only)
+  --json-output                    emit {"templates": [...]} instead of a table
+"""
+
+from __future__ import annotations
+
+import argparse
+import json
+import os
+import sys
+
+# When run as `python3 <skillDir>/scripts/list_templates.py`, this script's own
+# directory is sys.path[0], so the sibling shared modules import directly.
+from registry_loader import list_templates as load_visible_templates  # noqa: E402
+
+
+def _matches(tpl: dict, tag: str | None, aspect: str | None, language: str | None) -> bool:
+    if tag:
+        needle = tag.lower()
+        tags = [str(t).lower() for t in (tpl.get("styleTags") or [])]
+        if not any(needle in t or t in needle for t in tags):
+            return False
+    if aspect:
+        want = aspect.lower()
+        ratios = [str(r).lower() for r in (tpl.get("supportedAspectRatios") or [])]
+        if want not in ratios:
+            return False
+    if language:
+        # Unknown language (absent/empty) is never excluded by a language filter.
+        langs = [str(l).lower() for l in (tpl.get("contentLanguage") or [])]
+        if langs and language.lower() not in langs:
+            return False
+    return True
+
+
+def _status_of(tpl: dict) -> str:
+    raw = tpl.get("status")
+    if isinstance(raw, str) and raw.strip():
+        return raw.strip().lower()
+    return "stable"
+
+
+def main() -> None:
+    ap = argparse.ArgumentParser(
+        description="List available video templates from the registry.",
+    )
+    ap.add_argument("--list-templates", action="store_true",
+                    help="List every available template (the only CLI verb).")
+    ap.add_argument("--filter-tag", help="Keep templates whose styleTags match this substring.")
+    ap.add_argument("--filter-aspect", help="Keep templates supporting this aspect ratio (e.g. 9:16).")
+    ap.add_argument("--filter-language", help="Keep templates tagged with this content language (zh|en|...).")
+    ap.add_argument("--include-beta", action="store_true", help="Also show beta templates (default: stable only).")
+    ap.add_argument("--json-output", action="store_true", help='Emit {"templates": [...]} JSON instead of a table.')
+    args = ap.parse_args()
+
+    # --include-beta flips registry_loader's documented status gate via env, so
+    # the CLI and in-process callers share one gating implementation.
+    if args.include_beta:
+        os.environ["ENABLE_BETA_TEMPLATES"] = "1"
+
+    try:
+        templates = load_visible_templates()
+    except RuntimeError as exc:
+        print(f"❌ {exc}", file=sys.stderr)
+        sys.exit(1)
+
+    visible = [
+        t for t in templates
+        if _matches(t, args.filter_tag, args.filter_aspect, args.filter_language)
+    ]
+
+    if args.json_output:
+        print(json.dumps({"templates": visible}, ensure_ascii=False))
+        return
+
+    if not visible:
+        print(
+            "⚠️  No templates match the current filters "
+            "(check --include-beta or relax --filter-*).",
+        )
+        return
+
+    header = (
+        f"\n{'Template ID':<22} {'Name':<18} {'Aspect':<12} "
+        f"{'Lang':<8} {'Status':<8} Tags / LLM hint"
+    )
+    print(header)
+    print(
+        f"{'-' * 22} {'-' * 18} {'-' * 12} {'-' * 8} {'-' * 8} {'-' * 60}"
+    )
+    for tpl in visible:
+        tid = str(tpl.get("templateId", ""))[:22].ljust(22)
+        name = str(tpl.get("name", ""))[:18].ljust(18)
+        ratios = ",".join(tpl.get("supportedAspectRatios") or [])[:12].ljust(12)
+        lang = (",".join(tpl.get("contentLanguage") or []) or "-")[:8].ljust(8)
+        status = _status_of(tpl).ljust(8)
+        tags = ", ".join((tpl.get("styleTags") or [])[:5])
+        print(f"{tid} {name} {ratios} {lang} {status} {tags}")
+        hint = tpl.get("llmHint")
+        if hint:
+            hint = hint if len(hint) <= 200 else hint[:197] + "..."
+            print(f"{' ' * 22}   ↳ {hint}")
+
+    print(f"\n{len(visible)} template(s) shown")
+
+
+if __name__ == "__main__":
+    main()

package/skills/{template-bind → template-registry}/scripts/match_template.py

@@ -1,10 +1,11 @@
 """
-template-bind 库模块 — DSL → TemplateBinding 的纯函数实现。
+template-registry 库模块 — DSL → TemplateBinding 的纯函数实现。
 
 仅作为 Python library 被 render-video 的 render_video.py / gen-script 的
 gen_script.py 通过 `sys.path.insert` 后 `import match_template` 引入；
-对外不再暴露 CLI（template-bind skill 已迁移到 ab-skill CLI 的 builtin
-handler，由 src/handlers/template-bind.ts 处理 --list-templates）。
+对外不暴露 binding CLI。template-registry skill 的 `--list-templates` 入口由
+sibling 的 `list_templates.py` 实现（它复用 registry_loader 的统一加载 +
+状态门控），两者共享同一份 registry 读取逻辑，无 TS 侧重复实现。
 
 提供的公共接口：
   load_registry()                 → list[dict]   模板列表

package/skills/{template-bind → template-registry}/scripts/registry_loader.py

@@ -3,7 +3,7 @@
 模板 registry 加载器 - 统一抽象本地文件 / HTTP API 两种数据源。
 
 为什么独立成一个模块：
-  template-bind、render-video 等多个 skill 都要读 registry，集中在一处便于：
+  template-registry、render-video 等多个 skill 都要读 registry，集中在一处便于：
     - 缓存策略一致（HTTP 拉取 5 分钟 TTL + ETag 复用）
     - 解析顺序一致（环境变量 → 本地文件 → monorepo 回退）
     - 错误降级一致（HTTP 故障时复用上一次缓存）

package/skills/{template-bind → template-registry}/scripts/template_paths.py

@@ -6,9 +6,9 @@ Before this file landed, every caller that needed to reach into the
 monorepo's ``template-library/`` from inside ``apps/ab-skill/`` had to count
 ``..`` segments by hand. The depth differs per call site:
 
-    skills/template-bind/scripts/registry_loader.py        → 5 × ".."
-    skills/template-bind/scripts/check_contracts.py        → 3 × ".."
-    skills/template-bind/video_dsl/runtime/template_binder → 6 × ".."
+    skills/template-registry/scripts/registry_loader.py    → 5 × ".."
+    skills/template-registry/scripts/check_contracts.py    → 3 × ".."
+    skills/template-registry/video_dsl/runtime/template_binder → 6 × ".."
     skills/gen-script/scripts/gen_script.py                → 5 × ".."
     skills/render-video/scripts/render_video.py            → 4 × ".."
 
@@ -17,9 +17,9 @@ move silently breaks one site without touching the others. This module
 computes the monorepo root **once** from its own location and exposes
 helpers so callers no longer count.
 
-All sibling skills now go through these helpers; the TS handler in
-``src/handlers/template-bind.ts`` keeps its own one-off computation
-because it cannot import this Python module.
+All sibling skills (including the ``list_templates.py`` CLI entry, via
+``registry_loader``) go through these helpers, so the monorepo-root
+computation lives in exactly one place.
 """
 from __future__ import annotations
 
@@ -29,7 +29,7 @@ from typing import Optional
 __all__ = ["monorepo_root", "monorepo_registry_path", "monorepo_template_src_dir"]
 
 # This file lives at:
-#   <repo>/apps/ab-skill/skills/template-bind/scripts/template_paths.py
+#   <repo>/apps/ab-skill/skills/template-registry/scripts/template_paths.py
 # So the monorepo root is 5 levels up.
 _HERE = os.path.dirname(os.path.abspath(__file__))
 _REPO_ROOT = os.path.normpath(os.path.join(_HERE, "..", "..", "..", "..", ".."))

package/skills/{template-bind → template-registry}/skill.json

@@ -1,11 +1,11 @@
 {
-  "name": "template-bind",
-  "toolName": "template_bind",
+  "name": "template-registry",
+  "toolName": "template_registry",
   "tier": "orchestration",
-  "title": "Template Registry (List-only)",
+  "title": "Video Template Registry",
   "description": "List all available video templates (templateId / name / aspect ratio / style tags). Template-to-DSL binding is no longer exposed as a separate step — once prepare_video_assets receives a template_id it runs the full DSL→RenderPlan pipeline internally.",
   "envVars": ["PRIV_TOKEN", "VIDEO_TEMPLATE_REGISTRY", "VIDEO_TEMPLATE_REGISTRY_URL", "VIDEO_TEMPLATE_REGISTRY_HTTP_METHOD"],
-  "entry": { "type": "builtin", "handler": "template-bind" },
+  "entry": { "type": "python", "scriptPath": "scripts/list_templates.py" },
   "parameters": {
     "type": "object",
     "properties": {
@@ -13,7 +13,8 @@
       "filter_tag": { "type": "string", "description": "Show only templates whose styleTags contain (or are contained in) this string. Case-insensitive. E.g. 'tech' / '科普' / 'walkthrough'." },
       "filter_aspect": { "type": "string", "description": "Show only templates that declare this aspect ratio. E.g. '9:16' / '16:9' / '1:1'." },
       "filter_language": { "type": "string", "description": "Show only templates whose contentLanguage includes this code ('zh' or 'en'). Templates with no declared language are always shown (treated as language-agnostic)." },
-      "include_beta": { "type": "boolean", "description": "Also list templates with status='beta'. Default off (only 'stable' shows). The ENABLE_BETA_TEMPLATES env var has the same effect process-wide." }
+      "include_beta": { "type": "boolean", "description": "Also list templates with status='beta'. Default off (only 'stable' shows). The ENABLE_BETA_TEMPLATES env var has the same effect process-wide." },
+      "json_output": { "type": "boolean", "description": "Emit a JSON result ({ templates: [...] }) instead of the human-readable table." }
     },
     "required": []
   }

package/skills/template-registry/version.json

@@ -0,0 +1,7 @@
+{
+  "skillName": "template-registry",
+  "repoName": "agent-skill-media-maker",
+  "skillId": "475",
+  "version": "V12",
+  "skillDescription": "视频模板仓库（列表查询）。存储所有视频模板定义，对外只暴露「列出可用模板」一个能力；DSL→TemplateBinding 的绑定逻辑已内嵌进 prepare-video-assets，不再作为独立步骤暴露。同时是跨 skill 共享 Python 库（registry_loader / match_template / template_paths / video_dsl 等）的存放位置。\n\n当用户提到以下任何需求时，立即使用本 skill：\n- 查看可用模板、列出所有模板"
+}

package/skills/{template-bind → template-registry}/video_dsl/README.md

@@ -47,7 +47,7 @@ video_dsl/
 ```
 主题 → [gen-script] → Video DSL
                           ↓
-                   [template-bind] → TemplateBinding
+                   [template-registry] → TemplateBinding
                           ↓
                    [render-video] → RenderPlan → MP4 / 剪映草稿
 ```
@@ -67,7 +67,7 @@ video_dsl/
 | 模块 | 职责 | 被谁调用 |
 |------|------|----------|
 | `dsl_validator.py` | DSL JSON Schema 校验与参数标准化 | gen-script、render-video |
-| `template_binder.py` | 将 DSL scene 映射到模板 slot，生成 TemplateBinding（从 @ab-templates/metadata 读取模板定义） | template-bind |
+| `template_binder.py` | 将 DSL scene 映射到模板 slot，生成 TemplateBinding（从 @ab-templates/metadata 读取模板定义） | template-registry |
 | `timeline_compiler.py` | 帧号计算、字幕切片（`split_subtitle`）、旁白分段（`segment_narration`） | render-video（跨 skill import） |
 | `asset_adapter.py` | 素材格式适配 | render-video |
 | `prompt_enhancer.py` | 图片 prompt 增强 | render-video |

package/skills/{template-bind → template-registry}/video_dsl/runtime/__init__.py

@@ -1,7 +1,7 @@
 """Package marker for ``video_dsl.runtime`` — kept intentionally empty.
 
 This file exists so ``from video_dsl.runtime.<mod> import …`` resolves when
-``skills/template-bind/`` is on sys.path (the pattern used by render-video
+``skills/template-registry/`` is on sys.path (the pattern used by render-video
 and gen-script). Public modules:
 
     - timeline_compiler  : DSL → Remotion frame-axis timeline + subtitle splits