AIPing Media

AIPing 媒体生成技能，支持图片生成和视频生成。生成后自动下载到本地，优先通过飞书发送原生图片/视频消息，CDN 链接作为备选。当用户要求生成图片、生成视频、图片生成、视频制作、AI绘图、文生图、文生视频时自动触发。

Security Scan

The skill's code, required environment variables, and runtime instructions are consistent with its stated purpose (generate media via AIPing and send via Feishu); nothing requests unrelated credentials or installs arbitrary code.

Purpose & Capability

The skill name/description (AIPing media generation + send via Feishu) matches the required environment variables (AIPING_API_KEY, FEISHU_APP_ID, FEISHU_APP_SECRET) and the included scripts call AIPing and Feishu endpoints. No unrelated credentials or binaries are requested.

Instruction Scope

SKILL.md instructs setting env vars in ~/.openclaw/openclaw.json and restarting the gateway so the skill can read them — this is expected for a skill that needs API keys. The sample shell/python snippets read the local openclaw config (user-facing setup steps). The runtime scripts themselves read secrets from environment variables only and restrict local file reads to /tmp/. Minor issues: a sample shell snippet that reads the JSON path is somewhat awkward but not malicious.

Install Mechanism

No install spec; this is an instruction-only skill with included Python scripts. No downloads from untrusted URLs or archive extraction. The skill will run the provided scripts from the workspace.

Credentials

Requested environment variables are proportional and required for functionality: AIPING_API_KEY for the AIPing API, FEISHU_APP_ID/FEISHU_APP_SECRET for obtaining Feishu tenant token and uploading messages. No extra or unrelated secrets are requested.

Persistence & Privilege

always is false and the skill does not request permanent platform-wide privileges. It asks the user to store three env vars in the OpenClaw config (expected for API keys). The skill does not modify other skills' configs or system-wide settings beyond advising a gateway restart to pick up env changes.

Assessment

This skill appears to do what it says: it calls https://aiping.cn to generate media, downloads generated files into /tmp/, and uploads them to Feishu using the provided FEISHU_APP_ID/FEISHU_APP_SECRET. Before installing: 1) Only provide AIPING_API_KEY and Feishu app credentials you trust — these keys give the skill the ability to create media and send messages from the configured Feishu app. 2) Review the Feishu app permissions (im:resource and messaging scopes) and ensure the app is configured for the desired recipient scope. 3) Keep the keys secret in ~/.openclaw/openclaw.json and consider testing in an isolated environment first (the scripts download content to /tmp/ and will upload whatever is there). 4) Note a minor coding issue: send_feishu.py has a small typo in env lookup (FEESHU_APP_SECRET) but it falls back to the module-level variable; this looks like sloppy coding rather than malicious behavior. 5) If you need stricter controls, avoid storing credentials in global config and instead inject them at runtime or use scoped credentials. If you want me to, I can point out exact lines for the env-lookup typo and suggest small fixes to harden the scripts.

AIPing Media Skill

通过 AIPing API 生成图片和视频，完整流程：生成 → 本地下载 → 飞书原生发送，CDN 链接作为备选。

完整工作流

生成图片/视频 → 自动下载到本地 (/tmp/) → 上传飞书 → 发送原生消息
                                              ↓ 失败
                                         返回 CDN 链接

---

快速上手（新用户从零到跑通）

第一步：安装技能

clawhub install aiping-media

第二步：配置环境变量

在 ~/.openclaw/openclaw.json 的 env 块中添加三个变量。推荐用 CLI 写入，避免手动编辑 JSON：

openclaw config set env.AIPING_API_KEY   "你的AIPing Key"
openclaw config set env.FEISHU_APP_ID    "cli_xxxxxxxxxxxxxxxx"
openclaw config set env.FEISHU_APP_SECRET "你的飞书AppSecret"

获取飞书 AppID / AppSecret：飞书开放平台 → 你的应用 → 凭证与基础信息

验证写入成功：

openclaw config get env.AIPING_API_KEY
openclaw config get env.FEISHU_APP_ID

第三步：重启 Gateway（让 env 变量生效）

openclaw gateway restart

第四步：获取自己的飞书 open_id

以下命令自动用配置中的 AppID/Secret 获取 token，再查询你自己的 open_id（把 你的手机号或邮箱 替换成实际值）：

# 获取 token
TOKEN=$(python3 -c "
import requests, json
with open('$(echo ~/.openclaw/openclaw.json)') as f:
    cfg = json.load(f)
env = cfg['env']
r = requests.post('https://open.feishu.cn/open-apis/auth/v3/tenant_access_token/internal',
    json={'app_id': env['FEISHU_APP_ID'], 'app_secret': env['FEISHU_APP_SECRET']})
print(r.json()['tenant_access_token'])
")

# 按手机号查 open_id（换成你的手机号，含国家码如 [phone]）
curl -s "https://open.feishu.cn/open-apis/contact/v3/users/batch_get_id?user_id_type=open_id" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"mobiles": ["你的手机号"]}' | python3 -m json.tool

记下 open_id（格式如 ou_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx），后续测试用。

第五步：一键测试（图片生成 + 飞书发送）

SKILL_DIR=~/.openclaw/workspace/skills/aiping-media

# 生成图片（默认模型 Doubao-Seedream-5.0-lite，~5秒）
IMG_JSON=$(python3 $SKILL_DIR/scripts/generate_image.py "赛博朋克城市夜景，霓虹灯，超写实")
LOCAL_PATH=$(echo $IMG_JSON | python3 -c "import sys,json; print(json.load(sys.stdin)['local_path'])")
echo "图片已下载到: $LOCAL_PATH"

# 发送到飞书（替换 open_id）
python3 $SKILL_DIR/scripts/send_feishu.py image ou_你的open_id $LOCAL_PATH

输出 "success": true 即代表完整链路跑通。

第六步：测试视频生成

SKILL_DIR=~/.openclaw/workspace/skills/aiping-media

# 提交任务（视频生成通常需要 1~3 分钟）
TASK_JSON=$(python3 $SKILL_DIR/scripts/generate_video.py submit Kling-V3-Omni "一只橘猫在草地上打滚，阳光明媚" 5 pro 16:9)
TASK_ID=$(echo $TASK_JSON | python3 -c "import sys,json; print(json.load(sys.stdin)['id'])")
echo "Task ID: $TASK_ID"

# 等待完成并自动下载（轮询，最长等 5 分钟）
VIDEO_JSON=$(python3 $SKILL_DIR/scripts/generate_video.py wait $TASK_ID)
VIDEO_PATH=$(echo $VIDEO_JSON | python3 -c "import sys,json; print(json.load(sys.stdin).get('local_path',''))")
echo "视频已下载到: $VIDEO_PATH"

# 发送到飞书（群聊用 chat_id）
python3 $SKILL_DIR/scripts/send_feishu.py video oc_群聊ID $VIDEO_PATH chat_id

---

在 Agent 代码中调用

图片生成 + 飞书发送

from scripts.generate_image import generate_image
from scripts.send_feishu import send_image

# 1. 生成（自动下载到本地）
img = generate_image(
    model="Kling-V1",
    prompt="未来城市夜景，赛博朋克风格",
    negative_prompt="低质量，模糊"
)
local_path = img.get("local_path")  # 自动下载路径

# 2. 发送飞书
if local_path:
    # 发送给用户（默认 open_id）
    result = send_image(token=None, receive_id="ou_xxx", source=local_path)
    # 或发送到群聊（指定 chat_id）
    result = send_image(token=None, receive_id="oc_xxx", source=local_path, receive_id_type="chat_id")
    if result.get("success"):
        print(f"飞书 image_key: {result.get('image_key')}")
    else:
        print(f"失败: {result.get('error')}")
        print(f"CDN链接: {result.get('cdn_url')}")

视频生成 + 飞书发送

from scripts.generate_video import generate_video, wait_for_video
from scripts.send_feishu import send_video

# 1. 提交任务
video = generate_video(
    model="Kling-V3-Omni",
    prompt="一只猫咪在草地上玩耍",
    seconds=5,
    mode="pro",
    aspect_ratio="1:1"
)
task_id = video["id"]

# 2. 等待完成（自动下载）
result = wait_for_video(task_id)
local_path = result.get("local_path")

# 3. 发送到群聊
if local_path:
    result = send_video(token=None, receive_id="oc_xxx", source=local_path, receive_id_type="chat_id")

---

推荐模型

图片

模型	说明
Doubao-Seedream-5.0-lite	通用文生图，支持垫图（默认）
Kling-V1	高质量，速度快
HunyuanImage-3.0	腾讯混元

视频

模型	mode	aspect_ratio	说明
Kling-V3-Omni	pro/std	16:9, 9:16, 1:1	最新全能版，推荐
Kling-V2.6	std	16:9, 9:16, 1:1	稳定版

---

脚本命令行用法

generate_image.py

python3 scripts/generate_image.py <prompt> [model] [negative_prompt] [image_url] [aspect_ratio] [nodownload]

generate_video.py

# 提交任务
python3 scripts/generate_video.py submit <model> <prompt> [seconds] [mode] [aspect_ratio]
# 等待完成（自动下载）
python3 scripts/generate_video.py wait <task_id>
# 查询状态
python3 scripts/generate_video.py query <task_id>

send_feishu.py

# 图片（本地文件或 CDN URL）
python3 scripts/send_feishu.py image <receive_id> <file_or_url> [receive_id_type]
# 视频（本地文件或 CDN URL）
python3 scripts/send_feishu.py video <receive_id> <file_or_url> [receive_id_type]

关键参数 receive_id_type（默认 open_id，易踩坑！）：

receive_id_type	适用场景	receive_id 示例
open_id	发送给单个用户	ou_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
chat_id	发送到群聊	oc_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

⚠️ 群聊必须用 chat_id，不能拿用户的 open_id 发到群里，否则返回 400 Bad Request

示例：

# 发送给用户（默认 open_id）
python3 scripts/send_feishu.py image ou_xxxxxxxx /tmp/a.jpg

# 发送到群聊（指定 chat_id）
python3 scripts/send_feishu.py image oc_xxxxxxxx /tmp/a.jpg chat_id

---

飞书上传接口详解（踩坑经验汇总）

图片上传 vs 视频上传（对比）

	图片	视频
上传接口	POST /im/v1/images	POST /im/v1/files
参数传递方式	multipart/form-data（-F）	multipart/form-data（-F）
必需参数	image_type=message	file_type=mp4 + file_name=xxx.mp4
返回值 key	image_key	file_key
发送时 msg_type	image	media
content 格式	{"image_key": "xxx"}	{"file_key": "xxx", "type": "video"}

完整 curl 示例（图片）

# 1. 获取 token
TOKEN=$(curl -s -X POST "https://open.feishu.cn/open-apis/auth/v3/tenant_access_token/internal" \
  -H "Content-Type: application/json" \
  -d '{"app_id": "cli_xxx", "app_secret": "xxx"}' | python3 -c "import sys,json;print(json.load(sys.stdin)['tenant_access_token'])")

# 2. 上传图片
IMAGE_KEY=$(curl -s -X POST "https://open.feishu.cn/open-apis/im/v1/images" \
  -H "Authorization: Bearer $TOKEN" \
  -F "image_type=message" \
  -F "image=@/tmp/image.jpg" | python3 -c "import sys,json;print(json.load(sys.stdin)['data']['image_key'])")

# 3. 发送图片
curl -s -X POST "https://open.feishu.cn/open-apis/im/v1/messages?receive_id_type=chat_id" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d "{\"receive_id\": \"oc_xxx\", \"msg_type\": \"image\", \"content\": \"{\\"image_key\\": \\"$IMAGE_KEY\\"}\"}"

完整 curl 示例（视频）

# 1. 获取 token
TOKEN=$(curl -s -X POST "https://open.feishu.cn/open-apis/auth/v3/tenant_access_token/internal" \
  -H "Content-Type: application/json" \
  -d '{"app_id": "cli_xxx", "app_secret": "xxx"}' | python3 -c "import sys,json;print(json.load(sys.stdin)['tenant_access_token'])")

# 2. 上传视频（注意：file_type 必须是 "mp4"，不是 "video/mp4"！）
FILE_KEY=$(curl -s -X POST "https://open.feishu.cn/open-apis/im/v1/files" \
  -H "Authorization: Bearer $TOKEN" \
  -F "file_type=mp4" \
  -F "file_name=video.mp4" \
  -F "file=@/tmp/video.mp4" | python3 -c "import sys,json;print(json.load(sys.stdin)['data']['file_key'])")

# 3. 发送视频（msg_type 是 "media"，content 里 type 是 "video"）
curl -s -X POST "https://open.feishu.cn/open-apis/im/v1/messages?receive_id_type=chat_id" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d "{\"receive_id\": \"oc_xxx\", \"msg_type\": \"media\", \"content\": \"{\\"file_key\\": \\"$FILE_KEY\\", \\"type\\": \\"video\\"}\"}"

飞书上传最容易踩的坑

错误写法	正确写法	说明
上传图片设 Content-Type: application/json	不要设 Content-Type，让 curl 自动生成	multipart/form-data 不能手动设 Content-Type
file_type=video/mp4	file_type=mp4	文件类型是枚举值，不是 MIME type！
上传视频缺少 file_name	必须加 file_name=xxx.mp4	否则报 234001 参数错误
视频发送用 msg_type=video	用 msg_type=media	视频/音频/文件都是 media 类型
content: {"video_key": "xxx"}	content: {"file_key": "xxx", "type": "video"}	字段名是 file_key，不是 video_key

---

注意事项

1. env 变量必须配置：技能脚本从 AIPING_API_KEY / FEISHU_APP_ID / FEISHU_APP_SECRET 读取凭证，未配置时调用函数会抛 EnvironmentError
2. 重启 gateway 后才生效：修改 openclaw.json 的 env 块后必须 openclaw gateway restart
3. 本地文件限制 /tmp/：send_feishu.py 只允许读取 /tmp/ 下的本地文件，生成脚本默认也写入 /tmp/
4. 自动 MIME 检测：根据文件扩展名自动识别 PNG/JPEG/GIF/MP4
5. 视频生成需要耐心：视频任务提交后 status 为 queued，需要等待 1~3 分钟才能完成，用 wait 命令轮询查询
6. 上传接口不能混用：图片必须用 /im/v1/images，视频必须用 /im/v1/files，两者接口不同

常见错误排查

错误信息	原因	解决方案
400 Bad Request /im/v1/messages	receive_id_type 与 receive_id 不匹配（如群聊用了 open_id）	群聊传 chat_id + receive_id_type="chat_id"，个人用 open_id
飞书错误码 234001	参数错误，上传接口参数格式不对	对照上方「最容易踩的坑」检查：file_type 枚举值、file_name 是否传入、Content-Type 是否误设
飞书错误码 230001	缺少 im:resource 权限	飞书开放平台 → 权限管理 → 开通 im:resource → 重新发布
飞书错误码 99991663	tenant_access_token 过期	token 会自动续期，若仍失败重启 gateway
本地文件路径必须在 /tmp/	读取了 /tmp/ 以外的文件	只传 /tmp/ 下的文件路径
image_key / file_key 无效	图片上传和发送用了不同的 token 或应用	确保同一应用内完成上传和发送
视频发送成功但群里看不到	file_key 与 file_name 来自不同文件	每次发送都要重新上传，获取新的 file_key
AIPing 返回 503 暂无可用服务商	视频生成模型不支持 chat/completions 接口	视频生成必须用 /api/v1/videos 端点，不是 /chat/completions