{"id":"2ebe0697-0cc9-41ca-9f43-1fcf79b938cc","entityType":"agent","slug":"clawhub-xiaoyaocz-piccc-ai-skills","name":"Piccc AI Skills","canonicalUrl":"https://www.xpersona.co/agent/clawhub-xiaoyaocz-piccc-ai-skills","canonicalPath":"/agent/clawhub-xiaoyaocz-piccc-ai-skills","generatedAt":"2026-10-11T05:43:12.671Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-10T22:52:22.565Z","emptyReason":null},"description":"Generate images, videos, and audio with Piccc AI (皮可AI), or query and download existing media tasks. Use when the user asks to create visual or audio media with Piccc AI, discover its current models or voices, check an asynchronous task, or download completed results. Do not use for editing existing","descriptionLabel":"Source description","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 1.2K downloads reported by the source. Last updated 10/10/2026.","installCommand":"clawhub skill install s17d4nz97fhj3qj7erq9pnpe9d86hwph:piccc-ai-skills","sourceUrl":"https://clawhub.ai/xiaoyaocz/piccc-ai-skills","homepage":"https://clawhub.ai/xiaoyaocz/skills/piccc-ai-skills","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/xiaoyaocz/piccc-ai-skills","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/xiaoyaocz/skills/piccc-ai-skills","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":62,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"Piccc AI Skills technical dossier on Xpersona with agent coverage, OPENCLEW support, and live trust metadata."},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-10T22:52:22.565Z","emptyReason":null},"protocols":[{"protocol":"OPENCLEW","label":"OpenClaw","status":"self-declared","notes":"Declared in the public agent profile."}],"capabilities":[],"verifiedCount":0,"selfDeclaredCount":1,"capabilityMatrix":{"rows":[{"key":"OPENCLEW","type":"protocol","support":"unknown","confidenceSource":"profile","notes":"Listed on profile"}],"flattenedTokens":"protocol:OPENCLEW|unknown|profile"}},"adoption":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-10T22:52:22.565Z","emptyReason":null},"stars":null,"forks":null,"downloads":1237,"packageName":null,"latestVersion":"0.1.4","tractionLabel":"1.2K downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-10T22:52:22.506Z","emptyReason":null},"lastUpdatedAt":"2026-10-10T22:52:22.565Z","lastCrawledAt":"2026-10-10T22:52:22.506Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-11T22:52:22.506Z","lastVerifiedAt":null,"highlights":[{"version":"0.1.4","createdAt":"2026-07-28T09:29:49.290Z","changelog":"- Added onboarding and installation instructions for first-time users, including browser-based authorization workflow and balance reporting. - Introduced new language guidelines: always reply in the user's current language, and do not mix languages in responses (except for the product name). - Updated special-offer model warnings to use the user's language. - Clarified workflow and guardrails for authentication, credit retrieval, and command usage. - Added new auth command variant (`auth login --no-browser`) and direct credits query (`credits`) to the documented commands. - Added a banner image (assets/piccc-ai-banner.png) and removed skill-card.md from the package.","fileCount":13,"zipByteSize":6032485},{"version":"0.1.3","createdAt":"2026-07-28T08:47:42.952Z","changelog":"- Improved first-time authorization instructions: now clarifies browser-based flow and user actions if auto-launch fails. - Updated authentication workflow guidance to reflect the new automatic API Key saving and browser callback process. - Added explicit mention of the --no-browser auth option in the command list. - Removed obsolete skill-card.md; added a banner image (assets/piccc-ai-banner.png).","fileCount":13,"zipByteSize":6031522},{"version":"0.1.2","createdAt":"2026-07-11T13:49:32.483Z","changelog":"**Economy mode is now the default workflow, with stricter reliability and cost guardrails.** - Defaults to economy model and lowest-cost settings unless user requests higher quality or specific model. - Surfaces special-offer model reliability warnings (“特价模型可能生成较慢、稳定性较差”) before paid tasks. - Guardrails prevent raising quality, resolution, duration, or outputs above economy defaults without user input. - Clarified that only advertised model/voice IDs and parameters may be used; no inventing values. - Removed skill-card file; added banner asset.","fileCount":12,"zipByteSize":18897},{"version":"0.1.1","createdAt":"2026-07-11T12:08:36.497Z","changelog":"- Major authentication update: replaced manual API key handling with interactive authorization commands. - Added support for new auth commands: auth login, auth status, and auth logout. - Updated guardrails to prevent API key exposure and require script-managed credential storage. - Revised setup instructions for simpler, safer first-time use. - Removed mention of environment-variable setup unless explicitly requested by user.","fileCount":12,"zipByteSize":15849},{"version":"0.1.0","createdAt":"2026-07-11T11:15:33.991Z","changelog":"Initial release of Piccc AI skill. - Generate images, videos, and audio, or query and download tasks using Piccc AI. - Supports model/voice discovery and management of asynchronous tasks. - Enforces strict use of live model and voice lists; no invented parameters allowed. - Requires user to set the API key in the environment (never in chat). - Designed with guardrails to avoid extra charges, leaks, and premature success claims.","fileCount":4,"zipByteSize":4233}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s17d4nz97fhj3qj7erq9pnpe9d86hwph:piccc-ai-skills","setupComplexity":"low","setupSteps":["Install using `clawhub skill install s17d4nz97fhj3qj7erq9pnpe9d86hwph:piccc-ai-skills` in an isolated environment before connecting it to live workloads.","No published capability contract is available yet, so validate auth and request/response behavior manually.","Review the upstream CLAWHUB listing at https://clawhub.ai/xiaoyaocz/piccc-ai-skills before using production credentials."],"contract":{"contractStatus":"missing","authModes":[],"requires":[],"forbidden":[],"supportsMcp":false,"supportsA2a":false,"supportsStreaming":false,"inputSchemaRef":null,"outputSchemaRef":null,"dataRegion":null,"contractUpdatedAt":null,"sourceUpdatedAt":null,"freshnessSeconds":null},"invocationGuide":{"preferredApi":{"snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-xiaoyaocz-piccc-ai-skills/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-xiaoyaocz-piccc-ai-skills/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-xiaoyaocz-piccc-ai-skills/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-xiaoyaocz-piccc-ai-skills/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-xiaoyaocz-piccc-ai-skills/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-xiaoyaocz-piccc-ai-skills/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-11T05:43:12.668Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-xiaoyaocz-piccc-ai-skills/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-xiaoyaocz-piccc-ai-skills/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-xiaoyaocz-piccc-ai-skills/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-xiaoyaocz-piccc-ai-skills/trust"}},"reliability":{"evidence":{"source":"runtime-metrics","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No trust, reliability, or runtime telemetry is available."},"trust":{"status":"unavailable","handshakeStatus":"UNKNOWN","verificationFreshnessHours":null,"reputationScore":null,"p95LatencyMs":null,"successRate30d":null,"fallbackRate":null,"attempts30d":null,"trustUpdatedAt":null,"trustConfidence":"unknown","sourceUpdatedAt":null,"freshnessSeconds":null},"decisionGuardrails":{"doNotUseIf":["Contract metadata is missing or unavailable for deterministic execution."],"safeUseWhen":[],"riskFlags":["missing_or_unavailable_contract","trust_data_unavailable","schema_references_missing"],"operationalConfidence":"low"},"executionMetrics":{"observedLatencyMsP50":null,"observedLatencyMsP95":null,"estimatedCostUsd":null,"uptime30d":null,"rateLimitRpm":null,"rateLimitBurst":null,"lastVerifiedAt":null,"verificationSource":null},"runtimeMetrics":{"successRate":null,"avgLatencyMs":null,"avgCostUsd":null,"hallucinationRate":null,"retryRate":null,"disputeRate":null,"p50Latency":null,"p95Latency":null,"lastUpdated":null}},"benchmarks":{"evidence":{"source":"no-benchmark-data","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No benchmark suites or observed failure patterns are available."},"suites":[],"failurePatterns":[]},"artifacts":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-10T22:52:22.565Z","emptyReason":null},"readme":"Skill: Piccc AI Skills\n\nOwner: xiaoyaocz\n\nSummary: Generate images, videos, and audio with Piccc AI (皮可AI), or query and download existing media tasks. Use when the user asks to create visual or audio media with Piccc AI, discover its current models or voices, check an asynchronous task, or download completed results. Do not use for editing existing\n\nTags: latest:0.1.4\n\nVersion history:\n\nv0.1.4 | 2026-07-28T09:29:49.290Z | user\n\n- Added onboarding and installation instructions for first-time users, including browser-based authorization workflow and balance reporting.\n- Introduced new language guidelines: always reply in the user's current language, and do not mix languages in responses (except for the product name).\n- Updated special-offer model warnings to use the user's language.\n- Clarified workflow and guardrails for authentication, credit retrieval, and command usage.\n- Added new auth command variant (`auth login --no-browser`) and direct credits query (`credits`) to the documented commands.\n- Added a banner image (assets/piccc-ai-banner.png) and removed skill-card.md from the package.\n\nv0.1.3 | 2026-07-28T08:47:42.952Z | user\n\n- Improved first-time authorization instructions: now clarifies browser-based flow and user actions if auto-launch fails.\n- Updated authentication workflow guidance to reflect the new automatic API Key saving and browser callback process.\n- Added explicit mention of the --no-browser auth option in the command list.\n- Removed obsolete skill-card.md; added a banner image (assets/piccc-ai-banner.png).\n\nv0.1.2 | 2026-07-11T13:49:32.483Z | user\n\n**Economy mode is now the default workflow, with stricter reliability and cost guardrails.**\n\n- Defaults to economy model and lowest-cost settings unless user requests higher quality or specific model.\n- Surfaces special-offer model reliability warnings (“特价模型可能生成较慢、稳定性较差”) before paid tasks.\n- Guardrails prevent raising quality, resolution, duration, or outputs above economy defaults without user input.\n- Clarified that only advertised model/voice IDs and parameters may be used; no inventing values.\n- Removed skill-card file; added banner asset.\n\nv0.1.1 | 2026-07-11T12:08:36.497Z | user\n\n- Major authentication update: replaced manual API key handling with interactive authorization commands.\n- Added support for new auth commands: auth login, auth status, and auth logout.\n- Updated guardrails to prevent API key exposure and require script-managed credential storage.\n- Revised setup instructions for simpler, safer first-time use.\n- Removed mention of environment-variable setup unless explicitly requested by user.\n\nv0.1.0 | 2026-07-11T11:15:33.991Z | auto\n\nInitial release of Piccc AI skill.\n\n- Generate images, videos, and audio, or query and download tasks using Piccc AI.\n- Supports model/voice discovery and management of asynchronous tasks.\n- Enforces strict use of live model and voice lists; no invented parameters allowed.\n- Requires user to set the API key in the environment (never in chat).\n- Designed with guardrails to avoid extra charges, leaks, and premature success claims.\n\nArchive index:\n\nArchive v0.1.4: 13 files, 6032485 bytes\n\nFiles: .github/workflows/test.yml (268b), .gitignore (37b), agents/openai.yaml (211b), assets/piccc-ai-banner.png (6003805b), package.json (240b), README.md (2739b), references/api.md (2910b), scripts/piccc.mjs (31063b), skill-card.md (2147b), SKILL.md (5293b), test/skills.test.mjs (14847b), test/structure.test.mjs (2029b), _meta.json (134b)\n\nFile v0.1.4:SKILL.md\n\n---\nname: piccc-ai\ndescription: Generate images, videos, and audio with Piccc AI (皮可AI), or query and download existing media tasks. Use when the user asks to create visual or audio media with Piccc AI, discover its current models or voices, check an asynchronous task, or download completed results. Do not use for editing existing local media.\n---\n\n# Piccc AI (皮可AI)\n\nUse `{baseDir}/scripts/piccc.mjs` for every Piccc AI request. `{baseDir}` is the directory containing this file.\n\n## Language\n\n- Reply in the user's current language.\n- Do not mix translated Chinese and English phrases in one response. The product name `Piccc AI (皮可AI)` is the only intentional bilingual form.\n- Treat CLI output as internal data and explain it naturally in the user's language.\n\n## Installation and first authorization\n\nWhen the skill has just been installed, tell the user in their current language that installation is complete, browser authorization is required, and the available credit balance will be reported afterward.\n\nBefore the first request, run `node {baseDir}/scripts/piccc.mjs auth status`. If it reports `authenticated: false`:\n\n1. Explain that the user must sign in and approve access in the browser.\n2. Run `node {baseDir}/scripts/piccc.mjs auth login` and wait for it to finish. The command opens the authorization page, receives the local callback, saves the API key, and queries the available balance. Do not ask the user to report completion.\n3. If the browser cannot be opened, share the displayed link and verification code while the command keeps waiting. Never ask the user to create, copy, or paste an API key.\n4. After authorization, read `available_credits` from the result and report the balance in the user's current language.\n5. If `available_credits` is unavailable, run `node {baseDir}/scripts/piccc.mjs credits` and then report the returned balance.\n\n## Workflow\n\n1. Unless the user names a model or asks for higher quality, run `models TYPE --economy` and use its selected model and defaults. Prefer lower cost over speed or quality.\n2. Let explicit user requirements override economy defaults. Otherwise use one output, the lowest supported resolution and quality, the shortest supported video duration, no audio or web search, and the lowest supported audio sample rate.\n3. If `special_offer` is true, warn the user in their current language before submission that special-offer models may be slower and less reliable.\n4. Select only values advertised by the live model response. Never invent model IDs, voice IDs, formats, ratios, resolutions, modes, or durations.\n5. Put long or multiline prompts in a UTF-8 file and use `--prompt-file`.\n6. When the user expects a finished file, use `--wait --output-dir DIR` in the initial request. Do not create another task while waiting.\n7. On timeout, return the task ID and latest status so the same task can be resumed. Treat `completed`, `failed`, and `cancelled` as terminal.\n8. For audio, call the initial 10 credits a precharge and use `actual_cost` from the completed task as the final cost.\n\n## Commands\n\n```bash\nnode {baseDir}/scripts/piccc.mjs models image\nnode {baseDir}/scripts/piccc.mjs models video\nnode {baseDir}/scripts/piccc.mjs models audio\nnode {baseDir}/scripts/piccc.mjs models image --economy\nnode {baseDir}/scripts/piccc.mjs voices --model MODEL_ID --search Chinese\n\nnode {baseDir}/scripts/piccc.mjs generate image --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate video --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate audio --model MODEL_ID --prompt-file SCRIPT.txt --wait --output-dir OUTPUT_DIR\n\nnode {baseDir}/scripts/piccc.mjs task get TASK_ID\nnode {baseDir}/scripts/piccc.mjs task wait TASK_ID --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs tasks --type image --status completed --page 1 --page-size 20\nnode {baseDir}/scripts/piccc.mjs auth login\nnode {baseDir}/scripts/piccc.mjs auth login --no-browser\nnode {baseDir}/scripts/piccc.mjs auth status\nnode {baseDir}/scripts/piccc.mjs auth logout\nnode {baseDir}/scripts/piccc.mjs credits\nnode {baseDir}/scripts/piccc.mjs --help\n```\n\nRead [references/api.md](references/api.md) when selecting generation parameters or handling task responses.\n\n## Guardrails\n\n- Ask for clarification only when a missing creative choice would materially change the result.\n- Do not raise resolution, quality, duration, output count, audio, or web search above the economy defaults unless the user requests it or the task requires it.\n- Always surface the special-offer reliability warning before creating a paid task; a script warning is only a backup.\n- Never claim success before status is `completed` and at least one output URL exists.\n- Never expose the API key in command arguments, prompts, logs, or project files. Let the script manage its private credentials file.\n- Always use the authorization command for first-time setup; do not teach environment-variable setup unless the user explicitly requests it.\n- Do not create a replacement task while checking status; task reads must not trigger new charges.\n- Preserve the API response when reporting parameter, content-policy, billing, or upstream failures.\n\nFile v0.1.4:README.md\n\n# Piccc AI（皮可AI）技能\n\n[![皮可AI，让创意一键成型](assets/piccc-ai-banner.png)](https://picccai.cn)\n\n让 Codex、OpenClaw 和其他兼容智能体技能标准的工具调用 Piccc AI（皮可AI），生成图片、视频和音频，并自动等待、下载生成结果。\n\n## 快速安装\n\n把下面这句话发给智能体：\n\n```text\n安装技能 https://github.com/xiaoyaocz/piccc-ai-skills\n```\n\n也可以使用 Skills 命令行工具：\n\n```bash\nnpx skills add xiaoyaocz/piccc-ai-skills\n```\n\n安装程序会识别仓库中的 `piccc-ai` 技能，并让你选择 Codex、OpenClaw 或其他已支持的智能体。\n\n## 连接皮可AI账号\n\n安装完成后，智能体应主动提示你完成账号授权。也可以直接告诉智能体：\n\n```text\n连接我的皮可AI账号\n```\n\n智能体会自动打开授权页面。登录 Piccc AI（皮可AI）并确认授权后，页面会跳转到本地完成连接，智能体会自动继续、查询当前可用积分并告诉你；API 密钥会创建并保存在本机，不需要复制密钥、配置环境变量或再回复“已经完成”。\n\n如果浏览器无法自动打开，命令仍会显示授权链接和验证码，手动打开链接即可。也可以使用 `node scripts/piccc.mjs auth login --no-browser` 明确启用原来的手动授权方式。\n\n## 使用\n\n安装后直接告诉智能体你要生成什么，例如：\n\n```text\n用皮可AI生成一张 16:9 的电影感雪山日出图，完成后下载到 output/sunrise。\n```\n\n```text\n用皮可AI把这段产品介绍生成中文旁白，先列出可用音色让我选择。\n```\n\n```text\n查询皮可AI任务 task_xxx，完成后把结果下载到 output/piccc。\n```\n\n技能会先读取后台当前可用的模型和参数，不会固定模型名称。未指定规格时默认优先选择更优惠的模型和低分辨率、低质量配置；使用特价模型前会提示生成可能较慢、不稳定。\n\n查看当前授权状态：\n\n```bash\nnode scripts/piccc.mjs auth status\n```\n\n取消本机授权：\n\n```bash\nnode scripts/piccc.mjs auth logout\n```\n\n## 手动安装\n\n下载仓库后，把根目录中的 `SKILL.md`、`agents`、`references` 和 `scripts` 复制到对应的 `piccc-ai` 目录：\n\n| 智能体 | 当前项目 | 全局使用 |\n| --- | --- | --- |\n| Codex | `.agents/skills/piccc-ai` | `~/.codex/skills/piccc-ai` |\n| OpenClaw | `/skills/piccc-ai` | `~/.openclaw/skills/piccc-ai` |\n\nOpenClaw 也可以直接安装本地目录：\n\n```bash\nopenclaw skills install ./piccc-ai-skills\n```\n\n脚本需要 Node.js 20 或更高版本，不需要安装 npm 依赖。\n\n## 本地检查\n\n```bash\nnpm run check\n```\n\n查看脚本支持的命令：\n\n```bash\nnode scripts/piccc.mjs --help\n```\n\nFile v0.1.4:_meta.json\n\n{\n \"ownerId\": \"kn797z5rnyaaady2chsrh164ws86gp7h\",\n \"slug\": \"piccc-ai-skills\",\n \"version\": \"0.1.4\",\n \"publishedAt\": 1785230989290\n}\n\nFile v0.1.4:references/api.md\n\n# Piccc AI 媒体接口\n\n## 连接账号\n\n首次使用时运行：\n\n```bash\nnode scripts/piccc.mjs auth login\n```\n\n命令会自动打开授权页面。登录 Piccc AI（皮可AI）并确认授权后，页面会跳转到本机回调页，脚本随后自动创建并保存 API 密钥、查询当前可用积分。命令会一直等待授权结果，不要让用户在对话中发送密钥，也不要要求用户另行回复“已完成”。浏览器无法打开时，使用命令显示的链接和验证码；`auth login --no-browser` 可保留纯手动流程。\n\n## 模型接口\n\n| 类型 | 接口 | 主要能力字段 |\n| --- | --- | --- |\n| 图片 | `GET /v1/images/models` | 比例、分辨率、质量、生成数量 |\n| 视频 | `GET /v1/videos/models` | 分辨率、比例、时长、生成模式、参考模式 |\n| 音频 | `GET /v1/audio/models` | 格式、采样率、参数范围、参考模式、音色 |\n\n每次创建任务前使用模型响应中的 `data[].id`。音色使用 `voice_presets[].id`，不要传显示名称。\n\n## 经济规格\n\n默认先运行：\n\n```bash\nnode scripts/piccc.mjs models image --economy\n```\n\n将 `image` 换成 `video` 或 `audio`。命令会返回推荐模型、`defaults`、`special_offer` 和 `warning`。\n\n- 用户没有指定时，使用 `defaults` 中的低分辨率、低质量、单个输出、最短视频时长和低采样率。\n- 默认关闭视频音频、联网搜索和不必要的参考素材。\n- 用户明确要求高清、高质量、更长时长或指定模型时，以用户要求为准。\n- `special_offer=true` 时，创建任务前提示：“特价模型可能生成较慢、稳定性较差。”\n- 即使没有运行 `--economy`，生成脚本也会为未传参数补上当前模型支持的经济规格。\n\n## 创建任务\n\n图片：`POST /v1/images/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`aspect_ratio`、`resolution`、`quality`、`n`、`web_search`、`external_id`\n\n视频：`POST /v1/videos/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`route_mode`、`reference_mode`、`resolution`、`aspect_ratio`、`duration_seconds`、`audio`、`web_search`、`external_id`\n\n音频：`POST /v1/audio/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`output_format`、`sample_rate`、`speech_rate`、`loudness_rate`、`pitch_rate`、`reference_mode`、`voice_id`、`audio_references`、`image_references`、`external_id`\n- `reference_mode=audio` 时传 `voice_id` 或 `audio_references`\n- `reference_mode=image` 时传 `image_references`\n- 提交时预扣 10 积分，完成后的 `actual_cost` 是最终费用\n\n## 查询任务\n\n- `GET /v1/tasks/{task_id}`：查询单个图片、视频或音频任务。\n- `GET /v1/tasks`：分页查询任务，可按 `type` 和 `status` 过滤。\n\n任务状态包括 `queued`、`running`、`completed`、`failed`、`cancelled`。只有 `completed` 状态下的 `outputs[].url` 才是最终结果。\n\nFile v0.1.4:skill-card.md\n\n## Description:\n\nGenerate images, videos, and audio with Piccc AI (皮可AI), and query or download completed media tasks.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[xiaoyaocz](https://clawhub.ai/user/xiaoyaocz)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nDevelopers and agent users use this skill to authorize a Piccc AI account, discover available media models and voices, submit image, video, or audio generation tasks, and wait for or download completed outputs.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: Linking a Piccc AI account and storing a local API key can expose account tokens on shared or untrusted machines.\n\nMitigation: Install only from a reviewed source, keep the credentials file private, and run auth logout when finished on shared machines.\n\nRisk: Custom Piccc AI API or authorization base URLs can redirect credential or task traffic outside the intended service.\n\nMitigation: Avoid setting PICCC_API_BASE_URL or PICCC_AUTH_BASE_URL outside controlled testing.\n\nRisk: Generated media downloads may place files in user-selected locations.\n\nMitigation: Review output directories before downloading completed image, video, or audio results.\n\n## Reference(s):\n\n- [Piccc AI media API reference](references/api.md)\n- [Piccc AI website](https://picccai.cn)\n- [ClawHub skill page](https://clawhub.ai/xiaoyaocz/skills/piccc-ai-skills)\n\n## Skill Output:\n\n**Output Type(s):** [text, markdown, shell commands, files, guidance]\n\n**Output Format:** [Markdown responses with shell commands and downloaded media files when tasks complete]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [May produce downloaded image, video, or audio files after asynchronous Piccc AI tasks complete.]\n\n## Skill Version(s):\n\n0.1.4 (source: server release evidence; package.json reports 0.1.0)\n\n## Ethical Considerations:\n\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.\n\nFile v0.1.4:package.json\n\n{\n \"name\": \"piccc-ai-skills\",\n \"version\": \"0.1.0\",\n \"private\": true,\n \"type\": \"module\",\n \"engines\": {\n \"node\": \">=20\"\n },\n \"scripts\": {\n \"test\": \"node --test\",\n \"check\": \"node --check scripts/piccc.mjs && node --test\"\n }\n}\n\nFile v0.1.4:.github/workflows/test.yml\n\nname: Test skill\n\non:\n push:\n pull_request:\n\npermissions:\n contents: read\n\njobs:\n test:\n runs-on: ubuntu-latest\n steps:\n - uses: actions/checkout@v4\n - uses: actions/setup-node@v4\n with:\n node-version: 20\n - run: npm run check\n\nFile v0.1.4:agents/openai.yaml\n\ninterface:\n display_name: \"Piccc AI (皮可AI)\"\n short_description: \"Generate and download images, videos, and audio\"\n default_prompt: \"Use $piccc-ai to authorize my account and report my available credits.\"\n\nArchive v0.1.3: 13 files, 6031522 bytes\n\nFiles: .github/workflows/test.yml (268b), .gitignore (37b), agents/openai.yaml (249b), assets/piccc-ai-banner.png (6003805b), package.json (240b), README.md (2565b), references/api.md (2865b), scripts/piccc.mjs (29829b), skill-card.md (2063b), SKILL.md (4519b), test/skills.test.mjs (13844b), test/structure.test.mjs (1713b), _meta.json (134b)\n\nFile v0.1.3:SKILL.md\n\n---\nname: piccc-ai\ndescription: Generate images, videos, and audio with Piccc AI (皮可AI), or query and download existing media tasks. Use when the user asks to create visual or audio media with Piccc AI, discover its current models or voices, check an asynchronous task, or download completed results. Do not use for editing existing local media.\n---\n\n# Piccc AI\n\nUse `{baseDir}/scripts/piccc.mjs` for every Piccc AI request. `{baseDir}` is the directory containing this file.\n\n## Before the first request\n\nRun `node {baseDir}/scripts/piccc.mjs auth status`. If it reports `authenticated: false`, run `node {baseDir}/scripts/piccc.mjs auth login` and wait for the command to finish. The command opens the authorization page, receives the local browser callback, and saves the API Key automatically. Do not ask the user to report that authorization is complete. If the browser cannot be opened, give the user the displayed link and code while the command continues waiting. Never ask the user to create, copy, or paste an API Key.\n\n## Workflow\n\n1. Unless the user names a model or asks for higher quality, run `models TYPE --economy` and use its selected model and defaults. Prefer lower cost over speed or quality.\n2. Let explicit user requirements override economy defaults. Otherwise use one output, the lowest supported resolution and quality, the shortest supported video duration, no audio or web search, and the lowest supported audio sample rate.\n3. If `special_offer` is true or the selected model name or description says `特价`, `折扣`, `special offer`, `discount`, or `promo`, tell the user before submission: “特价模型可能生成较慢、稳定性较差。”\n4. Select only values advertised by the live model response. Never invent model IDs, voice IDs, formats, ratios, resolutions, modes, or durations.\n5. Put long or multiline prompts in a UTF-8 file and use `--prompt-file`.\n6. When the user expects a finished file, use `--wait --output-dir DIR` in the initial request. Do not create another task while waiting.\n7. On timeout, return the task ID and latest status so the same task can be resumed. Treat `completed`, `failed`, and `cancelled` as terminal.\n8. For audio, call the initial 10 credits a precharge and use `actual_cost` from the completed task as the final cost.\n\n## Commands\n\n```bash\nnode {baseDir}/scripts/piccc.mjs models image\nnode {baseDir}/scripts/piccc.mjs models video\nnode {baseDir}/scripts/piccc.mjs models audio\nnode {baseDir}/scripts/piccc.mjs models image --economy\nnode {baseDir}/scripts/piccc.mjs voices --model MODEL_ID --search Chinese\n\nnode {baseDir}/scripts/piccc.mjs generate image --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate video --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate audio --model MODEL_ID --prompt-file SCRIPT.txt --wait --output-dir OUTPUT_DIR\n\nnode {baseDir}/scripts/piccc.mjs task get TASK_ID\nnode {baseDir}/scripts/piccc.mjs task wait TASK_ID --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs tasks --type image --status completed --page 1 --page-size 20\nnode {baseDir}/scripts/piccc.mjs auth login\nnode {baseDir}/scripts/piccc.mjs auth login --no-browser\nnode {baseDir}/scripts/piccc.mjs auth status\nnode {baseDir}/scripts/piccc.mjs auth logout\nnode {baseDir}/scripts/piccc.mjs --help\n```\n\nRead [references/api.md](references/api.md) when selecting generation parameters or handling task responses.\n\n## Guardrails\n\n- Ask for clarification only when a missing creative choice would materially change the result.\n- Do not raise resolution, quality, duration, output count, audio, or web search above the economy defaults unless the user requests it or the task requires it.\n- Always surface the special-offer reliability warning before creating a paid task; a script warning is only a backup and does not replace telling the user.\n- Never claim success before status is `completed` and at least one output URL exists.\n- Never expose the API Key in command arguments, prompts, logs, or project files. Let the script manage its private credentials file.\n- Always use the authorization command for first-time setup; do not teach environment-variable setup unless the user explicitly requests it.\n- Do not create a replacement task while checking status; task reads must not trigger new charges.\n- Preserve the API response when reporting parameter, content-policy, billing, or upstream failures.\n\nFile v0.1.3:README.md\n\n# Piccc AI Skill\n\n[![皮可AI，让创意一键成型](assets/piccc-ai-banner.png)](https://picccai.cn)\n\n让 Codex、OpenClaw 和其他兼容 Agent Skills 的智能体调用皮可AI（Piccc AI），生成图片、视频和音频，并自动等待、下载生成结果。\n\n## 快速安装\n\n把下面这句话发给 Agent：\n\n```text\n安装技能 https://github.com/xiaoyaocz/piccc-ai-skills\n```\n\n也可以使用 Skills CLI：\n\n```bash\nnpx skills add xiaoyaocz/piccc-ai-skills\n```\n\n安装程序会识别仓库中的 `piccc-ai` 技能，并让你选择 Codex、OpenClaw 或其他已支持的 Agent。\n\n## 连接皮可AI账号\n\n安装后告诉 Agent：\n\n```text\n连接我的皮可AI账号\n```\n\nAgent 会自动打开授权页面。登录皮可AI并确认授权后，页面会跳转到本地完成连接，Agent 会自动继续；API Key 会创建并保存在本机，不需要复制密钥、配置环境变量或再告诉 Agent“已经完成”。\n\n如果浏览器无法自动打开，命令仍会显示授权链接和验证码，手动打开链接即可。也可以使用 `node scripts/piccc.mjs auth login --no-browser` 明确启用原来的手动授权方式。\n\n## 使用\n\n安装后直接告诉 Agent 你要生成什么，例如：\n\n```text\n用皮可AI生成一张 16:9 的电影感雪山日出图，完成后下载到 output/sunrise。\n```\n\n```text\n用皮可AI把这段产品介绍生成中文旁白，先列出可用音色让我选择。\n```\n\n```text\n查询皮可AI任务 task_xxx，完成后把结果下载到 output/piccc。\n```\n\n技能会先读取后台当前可用的模型和参数，不会固定模型名称。未指定规格时默认优先选择更优惠的模型和低分辨率、低质量配置；使用特价模型前会提示生成可能较慢、不稳定。\n\n查看当前授权状态：\n\n```bash\nnode scripts/piccc.mjs auth status\n```\n\n取消本机授权：\n\n```bash\nnode scripts/piccc.mjs auth logout\n```\n\n## 手动安装\n\n下载仓库后，把根目录中的 `SKILL.md`、`agents`、`references` 和 `scripts` 复制到对应的 `piccc-ai` 目录：\n\n| Agent | 当前项目 | 全局使用 |\n| --- | --- | --- |\n| Codex | `.agents/skills/piccc-ai` | `~/.codex/skills/piccc-ai` |\n| OpenClaw | `/skills/piccc-ai` | `~/.openclaw/skills/piccc-ai` |\n\nOpenClaw 也可以直接安装本地目录：\n\n```bash\nopenclaw skills install ./piccc-ai-skills\n```\n\n脚本需要 Node.js 20 或更高版本，不需要安装 npm 依赖。\n\n## 本地检查\n\n```bash\nnpm run check\n```\n\n查看脚本支持的命令：\n\n```bash\nnode scripts/piccc.mjs --help\n```\n\nFile v0.1.3:_meta.json\n\n{\n \"ownerId\": \"kn797z5rnyaaady2chsrh164ws86gp7h\",\n \"slug\": \"piccc-ai-skills\",\n \"version\": \"0.1.3\",\n \"publishedAt\": 1785228462952\n}\n\nFile v0.1.3:references/api.md\n\n# Piccc AI 媒体接口\n\n## 连接账号\n\n首次使用时运行：\n\n```bash\nnode scripts/piccc.mjs auth login\n```\n\n命令会自动打开授权页面。登录皮可AI并确认授权后，页面会跳转到本机回调页，脚本随后自动创建并保存 API Key。命令会一直等待授权结果，不要让用户在对话中发送密钥，也不要要求用户另行回复“已完成”。浏览器无法打开时，使用命令显示的链接和验证码；`auth login --no-browser` 可保留纯手动流程。\n\n## 模型接口\n\n| 类型 | 接口 | 主要能力字段 |\n| --- | --- | --- |\n| 图片 | `GET /v1/images/models` | 比例、分辨率、质量、生成数量 |\n| 视频 | `GET /v1/videos/models` | 分辨率、比例、时长、生成模式、参考模式 |\n| 音频 | `GET /v1/audio/models` | 格式、采样率、参数范围、参考模式、音色 |\n\n每次创建任务前使用模型响应中的 `data[].id`。音色使用 `voice_presets[].id`，不要传显示名称。\n\n## 经济规格\n\n默认先运行：\n\n```bash\nnode scripts/piccc.mjs models image --economy\n```\n\n将 `image` 换成 `video` 或 `audio`。命令会返回推荐模型、`defaults`、`special_offer` 和 `warning`。\n\n- 用户没有指定时，使用 `defaults` 中的低分辨率、低质量、单个输出、最短视频时长和低采样率。\n- 默认关闭视频音频、联网搜索和不必要的参考素材。\n- 用户明确要求高清、高质量、更长时长或指定模型时，以用户要求为准。\n- `special_offer=true` 时，创建任务前提示：“特价模型可能生成较慢、稳定性较差。”\n- 即使没有运行 `--economy`，生成脚本也会为未传参数补上当前模型支持的经济规格。\n\n## 创建任务\n\n图片：`POST /v1/images/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`aspect_ratio`、`resolution`、`quality`、`n`、`web_search`、`external_id`\n\n视频：`POST /v1/videos/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`route_mode`、`reference_mode`、`resolution`、`aspect_ratio`、`duration_seconds`、`audio`、`web_search`、`external_id`\n\n音频：`POST /v1/audio/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`output_format`、`sample_rate`、`speech_rate`、`loudness_rate`、`pitch_rate`、`reference_mode`、`voice_id`、`audio_references`、`image_references`、`external_id`\n- `reference_mode=audio` 时传 `voice_id` 或 `audio_references`\n- `reference_mode=image` 时传 `image_references`\n- 提交时预扣 10 积分，完成后的 `actual_cost` 是最终费用\n\n## 查询任务\n\n- `GET /v1/tasks/{task_id}`：查询单个图片、视频或音频任务。\n- `GET /v1/tasks`：分页查询任务，可按 `type` 和 `status` 过滤。\n\n任务状态包括 `queued`、`running`、`completed`、`failed`、`cancelled`。只有 `completed` 状态下的 `outputs[].url` 才是最终结果。\n\nFile v0.1.3:skill-card.md\n\n## Description:
\nGenerate images, videos, and audio with Piccc AI (皮可AI), or query and download existing media tasks.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[xiaoyaocz](https://clawhub.ai/user/xiaoyaocz)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers and agent users use this skill to connect a Piccc AI account, discover available media models and voices, create paid image, video, or audio generation tasks, monitor asynchronous task status, and download completed results.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: The skill can connect a Piccc AI account and create paid media-generation tasks.
\nMitigation: Review prompts, selected models, and generation settings before submission; use the skill's account logout command when local authorization should be removed.
\nRisk: Generated media and downloaded results come from an external media-generation service.
\nMitigation: Review completed outputs before relying on, publishing, or sharing them.
\n\n\n## Reference(s):
\n- [Piccc AI media API reference](references/api.md)
\n- [Piccc AI](https://picccai.cn)
\n- [ClawHub release page](https://clawhub.ai/xiaoyaocz/skills/piccc-ai-skills)
\n\n\n## Skill Output:
\n**Output Type(s):** [Guidance, Shell commands, Files, Configuration]
\n**Output Format:** [Markdown guidance with shell commands, task status summaries, and downloaded media file paths]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Uses live Piccc AI model responses for supported generation parameters and asynchronous task status.]
\n\n## Skill Version(s):
\n0.1.3 (source: server release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nFile v0.1.3:package.json\n\n{\n \"name\": \"piccc-ai-skills\",\n \"version\": \"0.1.0\",\n \"private\": true,\n \"type\": \"module\",\n \"engines\": {\n \"node\": \">=20\"\n },\n \"scripts\": {\n \"test\": \"node --test\",\n \"check\": \"node --check scripts/piccc.mjs && node --test\"\n }\n}\n\nFile v0.1.3:.github/workflows/test.yml\n\nname: Test skill\n\non:\n push:\n pull_request:\n\npermissions:\n contents: read\n\njobs:\n test:\n runs-on: ubuntu-latest\n steps:\n - uses: actions/checkout@v4\n - uses: actions/setup-node@v4\n with:\n node-version: 20\n - run: npm run check\n\nFile v0.1.3:agents/openai.yaml\n\ninterface:\n display_name: \"皮可AI（Piccc AI）\"\n short_description: \"使用皮可AI生成图片、视频和音频，并下载异步任务结果\"\n default_prompt: \"Use $piccc-ai to create media with Piccc AI and download the completed result.\"\n\nArchive v0.1.2: 12 files, 18897 bytes\n\nFiles: .github/workflows/test.yml (268b), .gitignore (37b), agents/openai.yaml (249b), package.json (240b), README.md (2273b), references/api.md (2608b), scripts/piccc.mjs (23541b), skill-card.md (2323b), SKILL.md (4243b), test/skills.test.mjs (12204b), test/structure.test.mjs (1713b), _meta.json (134b)\n\nFile v0.1.2:SKILL.md\n\n---\nname: piccc-ai\ndescription: Generate images, videos, and audio with Piccc AI (皮可AI), or query and download existing media tasks. Use when the user asks to create visual or audio media with Piccc AI, discover its current models or voices, check an asynchronous task, or download completed results. Do not use for editing existing local media.\n---\n\n# Piccc AI\n\nUse `{baseDir}/scripts/piccc.mjs` for every Piccc AI request. `{baseDir}` is the directory containing this file.\n\n## Before the first request\n\nRun `node {baseDir}/scripts/piccc.mjs auth status`. If it reports `authenticated: false`, run `node {baseDir}/scripts/piccc.mjs auth login`, give the user the displayed authorization link and verification code, and wait for the command to finish. Never ask the user to create, copy, or paste an API Key.\n\n## Workflow\n\n1. Unless the user names a model or asks for higher quality, run `models TYPE --economy` and use its selected model and defaults. Prefer lower cost over speed or quality.\n2. Let explicit user requirements override economy defaults. Otherwise use one output, the lowest supported resolution and quality, the shortest supported video duration, no audio or web search, and the lowest supported audio sample rate.\n3. If `special_offer` is true or the selected model name or description says `特价`, `折扣`, `special offer`, `discount`, or `promo`, tell the user before submission: “特价模型可能生成较慢、稳定性较差。”\n4. Select only values advertised by the live model response. Never invent model IDs, voice IDs, formats, ratios, resolutions, modes, or durations.\n5. Put long or multiline prompts in a UTF-8 file and use `--prompt-file`.\n6. When the user expects a finished file, use `--wait --output-dir DIR` in the initial request. Do not create another task while waiting.\n7. On timeout, return the task ID and latest status so the same task can be resumed. Treat `completed`, `failed`, and `cancelled` as terminal.\n8. For audio, call the initial 10 credits a precharge and use `actual_cost` from the completed task as the final cost.\n\n## Commands\n\n```bash\nnode {baseDir}/scripts/piccc.mjs models image\nnode {baseDir}/scripts/piccc.mjs models video\nnode {baseDir}/scripts/piccc.mjs models audio\nnode {baseDir}/scripts/piccc.mjs models image --economy\nnode {baseDir}/scripts/piccc.mjs voices --model MODEL_ID --search Chinese\n\nnode {baseDir}/scripts/piccc.mjs generate image --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate video --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate audio --model MODEL_ID --prompt-file SCRIPT.txt --wait --output-dir OUTPUT_DIR\n\nnode {baseDir}/scripts/piccc.mjs task get TASK_ID\nnode {baseDir}/scripts/piccc.mjs task wait TASK_ID --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs tasks --type image --status completed --page 1 --page-size 20\nnode {baseDir}/scripts/piccc.mjs auth login\nnode {baseDir}/scripts/piccc.mjs auth status\nnode {baseDir}/scripts/piccc.mjs auth logout\nnode {baseDir}/scripts/piccc.mjs --help\n```\n\nRead [references/api.md](references/api.md) when selecting generation parameters or handling task responses.\n\n## Guardrails\n\n- Ask for clarification only when a missing creative choice would materially change the result.\n- Do not raise resolution, quality, duration, output count, audio, or web search above the economy defaults unless the user requests it or the task requires it.\n- Always surface the special-offer reliability warning before creating a paid task; a script warning is only a backup and does not replace telling the user.\n- Never claim success before status is `completed` and at least one output URL exists.\n- Never expose the API Key in command arguments, prompts, logs, or project files. Let the script manage its private credentials file.\n- Always use the authorization command for first-time setup; do not teach environment-variable setup unless the user explicitly requests it.\n- Do not create a replacement task while checking status; task reads must not trigger new charges.\n- Preserve the API response when reporting parameter, content-policy, billing, or upstream failures.\n\nFile v0.1.2:README.md\n\n# Piccc AI Skill\n\n[![皮可AI，让创意一键成型](assets/piccc-ai-banner.png)](https://picccai.cn)\n\n让 Codex、OpenClaw 和其他兼容 Agent Skills 的智能体调用皮可AI（Piccc AI），生成图片、视频和音频，并自动等待、下载生成结果。\n\n## 快速安装\n\n把下面这句话发给 Agent：\n\n```text\n安装技能 https://github.com/xiaoyaocz/piccc-ai-skills\n```\n\n也可以使用 Skills CLI：\n\n```bash\nnpx skills add xiaoyaocz/piccc-ai-skills\n```\n\n安装程序会识别仓库中的 `piccc-ai` 技能，并让你选择 Codex、OpenClaw 或其他已支持的 Agent。\n\n## 连接皮可AI账号\n\n安装后告诉 Agent：\n\n```text\n连接我的皮可AI账号\n```\n\nAgent 会生成一个授权链接。打开链接，登录皮可AI并确认授权即可；API Key 会自动创建并保存在本机，不需要复制密钥或配置环境变量。\n\n## 使用\n\n安装后直接告诉 Agent 你要生成什么，例如：\n\n```text\n用皮可AI生成一张 16:9 的电影感雪山日出图，完成后下载到 output/sunrise。\n```\n\n```text\n用皮可AI把这段产品介绍生成中文旁白，先列出可用音色让我选择。\n```\n\n```text\n查询皮可AI任务 task_xxx，完成后把结果下载到 output/piccc。\n```\n\n技能会先读取后台当前可用的模型和参数，不会固定模型名称。未指定规格时默认优先选择更优惠的模型和低分辨率、低质量配置；使用特价模型前会提示生成可能较慢、不稳定。\n\n查看当前授权状态：\n\n```bash\nnode scripts/piccc.mjs auth status\n```\n\n取消本机授权：\n\n```bash\nnode scripts/piccc.mjs auth logout\n```\n\n## 手动安装\n\n下载仓库后，把根目录中的 `SKILL.md`、`agents`、`references` 和 `scripts` 复制到对应的 `piccc-ai` 目录：\n\n| Agent | 当前项目 | 全局使用 |\n| --- | --- | --- |\n| Codex | `.agents/skills/piccc-ai` | `~/.codex/skills/piccc-ai` |\n| OpenClaw | `/skills/piccc-ai` | `~/.openclaw/skills/piccc-ai` |\n\nOpenClaw 也可以直接安装本地目录：\n\n```bash\nopenclaw skills install ./piccc-ai-skills\n```\n\n脚本需要 Node.js 20 或更高版本，不需要安装 npm 依赖。\n\n## 本地检查\n\n```bash\nnpm run check\n```\n\n查看脚本支持的命令：\n\n```bash\nnode scripts/piccc.mjs --help\n```\n\nFile v0.1.2:_meta.json\n\n{\n \"ownerId\": \"kn797z5rnyaaady2chsrh164ws86gp7h\",\n \"slug\": \"piccc-ai-skills\",\n \"version\": \"0.1.2\",\n \"publishedAt\": 1783777772483\n}\n\nFile v0.1.2:references/api.md\n\n# Piccc AI 媒体接口\n\n## 连接账号\n\n首次使用时运行：\n\n```bash\nnode scripts/piccc.mjs auth login\n```\n\n打开命令显示的链接，登录皮可AI并确认授权。脚本会自动创建并保存 API Key，不要让用户在对话中发送密钥。\n\n## 模型接口\n\n| 类型 | 接口 | 主要能力字段 |\n| --- | --- | --- |\n| 图片 | `GET /v1/images/models` | 比例、分辨率、质量、生成数量 |\n| 视频 | `GET /v1/videos/models` | 分辨率、比例、时长、生成模式、参考模式 |\n| 音频 | `GET /v1/audio/models` | 格式、采样率、参数范围、参考模式、音色 |\n\n每次创建任务前使用模型响应中的 `data[].id`。音色使用 `voice_presets[].id`，不要传显示名称。\n\n## 经济规格\n\n默认先运行：\n\n```bash\nnode scripts/piccc.mjs models image --economy\n```\n\n将 `image` 换成 `video` 或 `audio`。命令会返回推荐模型、`defaults`、`special_offer` 和 `warning`。\n\n- 用户没有指定时，使用 `defaults` 中的低分辨率、低质量、单个输出、最短视频时长和低采样率。\n- 默认关闭视频音频、联网搜索和不必要的参考素材。\n- 用户明确要求高清、高质量、更长时长或指定模型时，以用户要求为准。\n- `special_offer=true` 时，创建任务前提示：“特价模型可能生成较慢、稳定性较差。”\n- 即使没有运行 `--economy`，生成脚本也会为未传参数补上当前模型支持的经济规格。\n\n## 创建任务\n\n图片：`POST /v1/images/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`aspect_ratio`、`resolution`、`quality`、`n`、`web_search`、`external_id`\n\n视频：`POST /v1/videos/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`route_mode`、`reference_mode`、`resolution`、`aspect_ratio`、`duration_seconds`、`audio`、`web_search`、`external_id`\n\n音频：`POST /v1/audio/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`output_format`、`sample_rate`、`speech_rate`、`loudness_rate`、`pitch_rate`、`reference_mode`、`voice_id`、`audio_references`、`image_references`、`external_id`\n- `reference_mode=audio` 时传 `voice_id` 或 `audio_references`\n- `reference_mode=image` 时传 `image_references`\n- 提交时预扣 10 积分，完成后的 `actual_cost` 是最终费用\n\n## 查询任务\n\n- `GET /v1/tasks/{task_id}`：查询单个图片、视频或音频任务。\n- `GET /v1/tasks`：分页查询任务，可按 `type` 和 `status` 过滤。\n\n任务状态包括 `queued`、`running`、`completed`、`failed`、`cancelled`。只有 `completed` 状态下的 `outputs[].url` 才是最终结果。\n\nFile v0.1.2:skill-card.md\n\n## Description:
\nGenerate images, videos, and audio with Piccc AI, or query and download existing media tasks.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[xiaoyaocz](https://clawhub.ai/user/xiaoyaocz)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nExternal users and developers use this skill to have an agent create Piccc AI image, video, and audio tasks, monitor asynchronous task status, and download completed outputs.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: The skill accesses a Piccc AI account and can create paid media-generation tasks.
\nMitigation: Install it only when that access is intended, review generation requests before submission, and use task-status commands rather than creating replacement tasks while waiting.
\nRisk: A local API key is saved after login.
\nMitigation: Use the provided logout command when access is no longer needed, or rotate the key in Piccc AI if the local environment is no longer trusted.
\nRisk: Generated media files are written to local storage.
\nMitigation: Download outputs into a dedicated folder selected for the task.
\nRisk: Special-offer models may be slower or less stable.
\nMitigation: Warn the user before creating a paid task when a selected model is marked as a special offer.
\n\n\n## Reference(s):
\n- [Piccc AI Media API](references/api.md)
\n- [Piccc AI](https://picccai.cn)
\n- [ClawHub Skill Page](https://clawhub.ai/xiaoyaocz/skills/piccc-ai-skills)
\n\n\n## Skill Output:
\n**Output Type(s):** [Text, Shell commands, JSON, Files, Guidance]
\n**Output Format:** [Markdown guidance with shell commands, JSON task responses, and downloaded media files]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Outputs may include generated image, video, or audio files downloaded to a user-selected directory.]
\n\n## Skill Version(s):
\n0.1.2 (source: ClawHub release metadata)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nFile v0.1.2:package.json\n\n{\n \"name\": \"piccc-ai-skills\",\n \"version\": \"0.1.0\",\n \"private\": true,\n \"type\": \"module\",\n \"engines\": {\n \"node\": \">=20\"\n },\n \"scripts\": {\n \"test\": \"node --test\",\n \"check\": \"node --check scripts/piccc.mjs && node --test\"\n }\n}\n\nFile v0.1.2:.github/workflows/test.yml\n\nname: Test skill\n\non:\n push:\n pull_request:\n\npermissions:\n contents: read\n\njobs:\n test:\n runs-on: ubuntu-latest\n steps:\n - uses: actions/checkout@v4\n - uses: actions/setup-node@v4\n with:\n node-version: 20\n - run: npm run check\n\nFile v0.1.2:agents/openai.yaml\n\ninterface:\n display_name: \"皮可AI（Piccc AI）\"\n short_description: \"使用皮可AI生成图片、视频和音频，并下载异步任务结果\"\n default_prompt: \"Use $piccc-ai to create media with Piccc AI and download the completed result.\"\n\nArchive v0.1.1: 12 files, 15849 bytes\n\nFiles: .github/workflows/test.yml (268b), .gitignore (37b), agents/openai.yaml (249b), package.json (240b), README.md (2120b), references/api.md (1871b), scripts/piccc.mjs (17485b), skill-card.md (2255b), SKILL.md (3348b), test/skills.test.mjs (10073b), test/structure.test.mjs (1713b), _meta.json (134b)\n\nFile v0.1.1:SKILL.md\n\n---\nname: piccc-ai\ndescription: Generate images, videos, and audio with Piccc AI (皮可AI), or query and download existing media tasks. Use when the user asks to create visual or audio media with Piccc AI, discover its current models or voices, check an asynchronous task, or download completed results. Do not use for editing existing local media.\n---\n\n# Piccc AI\n\nUse `{baseDir}/scripts/piccc.mjs` for every Piccc AI request. `{baseDir}` is the directory containing this file.\n\n## Before the first request\n\nRun `node {baseDir}/scripts/piccc.mjs auth status`. If it reports `authenticated: false`, run `node {baseDir}/scripts/piccc.mjs auth login`, give the user the displayed authorization link and verification code, and wait for the command to finish. Never ask the user to create, copy, or paste an API Key.\n\n## Workflow\n\n1. Query the relevant live model list before creating a task. Reuse a supplied model ID only when it came from a current response.\n2. Select only values advertised by that model. Never invent model IDs, voice IDs, formats, ratios, resolutions, modes, or durations.\n3. Put long or multiline prompts in a UTF-8 file and use `--prompt-file`.\n4. When the user expects a finished file, use `--wait --output-dir DIR` in the initial request. Do not create another task while waiting.\n5. On timeout, return the task ID and latest status so the same task can be resumed. Treat `completed`, `failed`, and `cancelled` as terminal.\n6. For audio, call the initial 10 credits a precharge and use `actual_cost` from the completed task as the final cost.\n\n## Commands\n\n```bash\nnode {baseDir}/scripts/piccc.mjs models image\nnode {baseDir}/scripts/piccc.mjs models video\nnode {baseDir}/scripts/piccc.mjs models audio\nnode {baseDir}/scripts/piccc.mjs voices --model MODEL_ID --search Chinese\n\nnode {baseDir}/scripts/piccc.mjs generate image --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate video --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate audio --model MODEL_ID --prompt-file SCRIPT.txt --wait --output-dir OUTPUT_DIR\n\nnode {baseDir}/scripts/piccc.mjs task get TASK_ID\nnode {baseDir}/scripts/piccc.mjs task wait TASK_ID --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs tasks --type image --status completed --page 1 --page-size 20\nnode {baseDir}/scripts/piccc.mjs auth login\nnode {baseDir}/scripts/piccc.mjs auth status\nnode {baseDir}/scripts/piccc.mjs auth logout\nnode {baseDir}/scripts/piccc.mjs --help\n```\n\nRead [references/api.md](references/api.md) when selecting generation parameters or handling task responses.\n\n## Guardrails\n\n- Ask for clarification only when a missing creative choice would materially change the result.\n- Never claim success before status is `completed` and at least one output URL exists.\n- Never expose the API Key in command arguments, prompts, logs, or project files. Let the script manage its private credentials file.\n- Always use the authorization command for first-time setup; do not teach environment-variable setup unless the user explicitly requests it.\n- Do not create a replacement task while checking status; task reads must not trigger new charges.\n- Preserve the API response when reporting parameter, content-policy, billing, or upstream failures.\n\nFile v0.1.1:README.md\n\n# Piccc AI Skill\n\n[![皮可AI，让创意一键成型](assets/piccc-ai-banner.png)](https://picccai.cn)\n\n让 Codex、OpenClaw 和其他兼容 Agent Skills 的智能体调用皮可AI（Piccc AI），生成图片、视频和音频，并自动等待、下载生成结果。\n\n## 快速安装\n\n把下面这句话发给 Agent：\n\n```text\n安装技能 https://github.com/xiaoyaocz/piccc-ai-skills\n```\n\n也可以使用 Skills CLI：\n\n```bash\nnpx skills add xiaoyaocz/piccc-ai-skills\n```\n\n安装程序会识别仓库中的 `piccc-ai` 技能，并让你选择 Codex、OpenClaw 或其他已支持的 Agent。\n\n## 连接皮可AI账号\n\n安装后告诉 Agent：\n\n```text\n连接我的皮可AI账号\n```\n\nAgent 会生成一个授权链接。打开链接，登录皮可AI并确认授权即可；API Key 会自动创建并保存在本机，不需要复制密钥或配置环境变量。\n\n## 使用\n\n安装后直接告诉 Agent 你要生成什么，例如：\n\n```text\n用皮可AI生成一张 16:9 的电影感雪山日出图，完成后下载到 output/sunrise。\n```\n\n```text\n用皮可AI把这段产品介绍生成中文旁白，先列出可用音色让我选择。\n```\n\n```text\n查询皮可AI任务 task_xxx，完成后把结果下载到 output/piccc。\n```\n\n技能会先读取后台当前可用的模型和参数，不会固定模型名称。\n\n查看当前授权状态：\n\n```bash\nnode scripts/piccc.mjs auth status\n```\n\n取消本机授权：\n\n```bash\nnode scripts/piccc.mjs auth logout\n```\n\n## 手动安装\n\n下载仓库后，把根目录中的 `SKILL.md`、`agents`、`references` 和 `scripts` 复制到对应的 `piccc-ai` 目录：\n\n| Agent | 当前项目 | 全局使用 |\n| --- | --- | --- |\n| Codex | `.agents/skills/piccc-ai` | `~/.codex/skills/piccc-ai` |\n| OpenClaw | `/skills/piccc-ai` | `~/.openclaw/skills/piccc-ai` |\n\nOpenClaw 也可以直接安装本地目录：\n\n```bash\nopenclaw skills install ./piccc-ai-skills\n```\n\n脚本需要 Node.js 20 或更高版本，不需要安装 npm 依赖。\n\n## 本地检查\n\n```bash\nnpm run check\n```\n\n查看脚本支持的命令：\n\n```bash\nnode scripts/piccc.mjs --help\n```\n\nFile v0.1.1:_meta.json\n\n{\n \"ownerId\": \"kn797z5rnyaaady2chsrh164ws86gp7h\",\n \"slug\": \"piccc-ai-skills\",\n \"version\": \"0.1.1\",\n \"publishedAt\": 1783771716497\n}\n\nFile v0.1.1:references/api.md\n\n# Piccc AI 媒体接口\n\n## 连接账号\n\n首次使用时运行：\n\n```bash\nnode scripts/piccc.mjs auth login\n```\n\n打开命令显示的链接，登录皮可AI并确认授权。脚本会自动创建并保存 API Key，不要让用户在对话中发送密钥。\n\n## 模型接口\n\n| 类型 | 接口 | 主要能力字段 |\n| --- | --- | --- |\n| 图片 | `GET /v1/images/models` | 比例、分辨率、质量、生成数量 |\n| 视频 | `GET /v1/videos/models` | 分辨率、比例、时长、生成模式、参考模式 |\n| 音频 | `GET /v1/audio/models` | 格式、采样率、参数范围、参考模式、音色 |\n\n每次创建任务前使用模型响应中的 `data[].id`。音色使用 `voice_presets[].id`，不要传显示名称。\n\n## 创建任务\n\n图片：`POST /v1/images/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`aspect_ratio`、`resolution`、`quality`、`n`、`web_search`、`external_id`\n\n视频：`POST /v1/videos/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`route_mode`、`reference_mode`、`resolution`、`aspect_ratio`、`duration_seconds`、`audio`、`web_search`、`external_id`\n\n音频：`POST /v1/audio/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`output_format`、`sample_rate`、`speech_rate`、`loudness_rate`、`pitch_rate`、`reference_mode`、`voice_id`、`audio_references`、`image_references`、`external_id`\n- `reference_mode=audio` 时传 `voice_id` 或 `audio_references`\n- `reference_mode=image` 时传 `image_references`\n- 提交时预扣 10 积分，完成后的 `actual_cost` 是最终费用\n\n## 查询任务\n\n- `GET /v1/tasks/{task_id}`：查询单个图片、视频或音频任务。\n- `GET /v1/tasks`：分页查询任务，可按 `type` 和 `status` 过滤。\n\n任务状态包括 `queued`、`running`、`completed`、`failed`、`cancelled`。只有 `completed` 状态下的 `outputs[].url` 才是最终结果。\n\nFile v0.1.1:skill-card.md\n\n## Description:
\nGenerate images, videos, and audio with Piccc AI, or query and download existing media tasks.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[xiaoyaocz](https://clawhub.ai/user/xiaoyaocz)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nExternal users and developers use this skill to connect a Piccc AI account, discover live Piccc models and voices, create image, video, or audio generation tasks, monitor task status, and download completed media.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: The skill requires authorizing Piccc AI on the local machine and keeps a local API key for future use.
\nMitigation: Use the provided auth status and auth logout commands to inspect or remove local authorization when it is no longer needed.
\nRisk: Prompts and task details are sent to Piccc AI for media generation.
\nMitigation: Avoid submitting sensitive or confidential content unless sharing it with Piccc AI is acceptable for the use case.
\nRisk: Generated media can be downloaded to paths selected during use.
\nMitigation: Choose explicit output directories and review downloaded files before using or redistributing them.
\n\n\n## Reference(s):
\n- [Piccc AI media API reference](references/api.md)
\n- [Piccc AI](https://picccai.cn)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, shell commands, configuration, guidance, files]
\n**Output Format:** [Markdown guidance with shell commands and JSON task responses; completed tasks may download image, video, or audio files.]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Uses asynchronous Piccc AI tasks; task status and output files depend on Piccc AI responses and user-selected download paths.]
\n\n## Skill Version(s):
\n0.1.1 (source: server release metadata; artifact package.json reports 0.1.0)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nFile v0.1.1:package.json\n\n{\n \"name\": \"piccc-ai-skills\",\n \"version\": \"0.1.0\",\n \"private\": true,\n \"type\": \"module\",\n \"engines\": {\n \"node\": \">=20\"\n },\n \"scripts\": {\n \"test\": \"node --test\",\n \"check\": \"node --check scripts/piccc.mjs && node --test\"\n }\n}\n\nFile v0.1.1:.github/workflows/test.yml\n\nname: Test skill\n\non:\n push:\n pull_request:\n\npermissions:\n contents: read\n\njobs:\n test:\n runs-on: ubuntu-latest\n steps:\n - uses: actions/checkout@v4\n - uses: actions/setup-node@v4\n with:\n node-version: 20\n - run: npm run check\n\nFile v0.1.1:agents/openai.yaml\n\ninterface:\n display_name: \"皮可AI（Piccc AI）\"\n short_description: \"使用皮可AI生成图片、视频和音频，并下载异步任务结果\"\n default_prompt: \"Use $piccc-ai to create media with Piccc AI and download the completed result.\"\n\nArchive v0.1.0: 4 files, 4233 bytes\n\nFiles: references/api.md (1886b), skill-card.md (2545b), SKILL.md (3074b), _meta.json (134b)\n\nFile v0.1.0:SKILL.md\n\n---\nname: piccc-ai\ndescription: Generate images, videos, and audio with Piccc AI (皮可AI), or query and download existing media tasks. Use when the user asks to create visual or audio media with Piccc AI, discover its current models or voices, check an asynchronous task, or download completed results. Do not use for editing existing local media.\n---\n\n# Piccc AI\n\nUse `{baseDir}/scripts/piccc.mjs` for every Piccc AI request. `{baseDir}` is the directory containing this file.\n\n## Before the first request\n\nCheck `PICCC_API_KEY`. If it is unavailable, direct the user to [皮可AI个人中心](https://picccai.cn/account?tab=apiKeys). Tell them to sign in, open “API Key,” create a key, and save it immediately because the complete key is shown only once. Never ask the user to paste the key into chat; have them store it in the Agent's environment or secret manager.\n\n## Workflow\n\n1. Query the relevant live model list before creating a task. Reuse a supplied model ID only when it came from a current response.\n2. Select only values advertised by that model. Never invent model IDs, voice IDs, formats, ratios, resolutions, modes, or durations.\n3. Put long or multiline prompts in a UTF-8 file and use `--prompt-file`.\n4. When the user expects a finished file, use `--wait --output-dir DIR` in the initial request. Do not create another task while waiting.\n5. On timeout, return the task ID and latest status so the same task can be resumed. Treat `completed`, `failed`, and `cancelled` as terminal.\n6. For audio, call the initial 10 credits a precharge and use `actual_cost` from the completed task as the final cost.\n\n## Commands\n\n```bash\nnode {baseDir}/scripts/piccc.mjs models image\nnode {baseDir}/scripts/piccc.mjs models video\nnode {baseDir}/scripts/piccc.mjs models audio\nnode {baseDir}/scripts/piccc.mjs voices --model MODEL_ID --search Chinese\n\nnode {baseDir}/scripts/piccc.mjs generate image --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate video --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate audio --model MODEL_ID --prompt-file SCRIPT.txt --wait --output-dir OUTPUT_DIR\n\nnode {baseDir}/scripts/piccc.mjs task get TASK_ID\nnode {baseDir}/scripts/piccc.mjs task wait TASK_ID --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs tasks --type image --status completed --page 1 --page-size 20\nnode {baseDir}/scripts/piccc.mjs --help\n```\n\nRead [references/api.md](references/api.md) when selecting generation parameters or handling task responses.\n\n## Guardrails\n\n- Ask for clarification only when a missing creative choice would materially change the result.\n- Never claim success before status is `completed` and at least one output URL exists.\n- Never expose the API Key in command arguments, prompts, logs, or files.\n- Do not create a replacement task while checking status; task reads must not trigger new charges.\n- Preserve the API response when reporting parameter, content-policy, billing, or upstream failures.\n\nFile v0.1.0:_meta.json\n\n{\n \"ownerId\": \"kn797z5rnyaaady2chsrh164ws86gp7h\",\n \"slug\": \"piccc-ai-skills\",\n \"version\": \"0.1.0\",\n \"publishedAt\": 1783768533991\n}\n\nFile v0.1.0:references/api.md\n\n# Piccc AI 媒体接口\n\n## 准备 API Key\n\n前往[皮可AI（Piccc AI）个人中心](https://picccai.cn/account?tab=apiKeys)，登录后进入“API Key”，创建并立即保存密钥。完整密钥只显示一次。\n\n脚本从 `PICCC_API_KEY` 环境变量读取密钥。\n\n## 模型接口\n\n| 类型 | 接口 | 主要能力字段 |\n| --- | --- | --- |\n| 图片 | `GET /v1/images/models` | 比例、分辨率、质量、生成数量 |\n| 视频 | `GET /v1/videos/models` | 分辨率、比例、时长、生成模式、参考模式 |\n| 音频 | `GET /v1/audio/models` | 格式、采样率、参数范围、参考模式、音色 |\n\n每次创建任务前使用模型响应中的 `data[].id`。音色使用 `voice_presets[].id`，不要传显示名称。\n\n## 创建任务\n\n图片：`POST /v1/images/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`aspect_ratio`、`resolution`、`quality`、`n`、`web_search`、`external_id`\n\n视频：`POST /v1/videos/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`route_mode`、`reference_mode`、`resolution`、`aspect_ratio`、`duration_seconds`、`audio`、`web_search`、`external_id`\n\n音频：`POST /v1/audio/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`output_format`、`sample_rate`、`speech_rate`、`loudness_rate`、`pitch_rate`、`reference_mode`、`voice_id`、`audio_references`、`image_references`、`external_id`\n- `reference_mode=audio` 时传 `voice_id` 或 `audio_references`\n- `reference_mode=image` 时传 `image_references`\n- 提交时预扣 10 积分，完成后的 `actual_cost` 是最终费用\n\n## 查询任务\n\n- `GET /v1/tasks/{task_id}`：查询单个图片、视频或音频任务。\n- `GET /v1/tasks`：分页查询任务，可按 `type` 和 `status` 过滤。\n\n任务状态包括 `queued`、`running`、`completed`、`failed`、`cancelled`。只有 `completed` 状态下的 `outputs[].url` 才是最终结果。\n\nFile v0.1.0:skill-card.md\n\n## Description:
\nGenerate images, videos, and audio with Piccc AI, query or download media tasks, discover current models or voices, and check asynchronous task status.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[xiaoyaocz](https://clawhub.ai/user/xiaoyaocz)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers and agents use this skill to create Piccc AI media tasks, discover supported models and voices, monitor asynchronous jobs, and download completed outputs while protecting API keys and avoiding duplicate charges.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Piccc AI task creation can spend account credits.
\nMitigation: Use live model and voice discovery before creating tasks, avoid duplicate replacement tasks while waiting, and report the completed task's final cost when available.
\nRisk: PICCC_API_KEY exposure could compromise the user's Piccc AI account.
\nMitigation: Keep the key in the agent environment or secret manager and never include it in chat, command arguments, prompts, logs, or files.
\nRisk: The submitted artifact references a helper script that is not included in the artifact file list.
\nMitigation: Verify the installed skill includes the expected helper script before relying on the skill for Piccc AI requests.
\n\n\n## Reference(s):
\n- [Piccc AI API reference](references/api.md)
\n- [ClawHub skill page](https://clawhub.ai/xiaoyaocz/skills/piccc-ai-skills)
\n- [Source repository (server-resolved GitHub import)](https://github.com/xiaoyaocz/piccc-ai-skills)
\n- [Piccc AI API key setup](https://picccai.cn/account?tab=apiKeys)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, shell commands, configuration, guidance, files]
\n**Output Format:** [Markdown guidance with inline shell commands and task/status summaries; downloaded media files may be saved to an output directory.]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Requires PICCC_API_KEY in the environment and uses live Piccc AI model and voice lists before task creation.]
\n\n## Skill Version(s):
\n0.1.0 (source: server release metadata)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
","readmeExcerpt":"Skill: Piccc AI Skills Owner: xiaoyaocz Summary: Generate images, videos, and audio with Piccc AI (皮可AI), or query and download existing media tasks. Use when the user asks to create visual or audio media with Piccc AI, discover its current models or voices, check an asynchronous task, or download completed results. Do not use for editing existing Tags: latest:0.1.4 Version history: v0.1.4 | 2026-07-28T09:29:49.290Z ","codeSnippets":[],"executableExamples":[{"language":"bash","snippet":"node {baseDir}/scripts/piccc.mjs models image\nnode {baseDir}/scripts/piccc.mjs models video\nnode {baseDir}/scripts/piccc.mjs models audio\nnode {baseDir}/scripts/piccc.mjs models image --economy\nnode {baseDir}/scripts/piccc.mjs voices --model MODEL_ID --search Chinese\n\nnode {baseDir}/scripts/piccc.mjs generate image --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate video --model MODEL_ID --prompt-file PROMPT.txt --wait --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs generate audio --model MODEL_ID --prompt-file SCRIPT.txt --wait --output-dir OUTPUT_DIR\n\nnode {baseDir}/scripts/piccc.mjs task get TASK_ID\nnode {baseDir}/scripts/piccc.mjs task wait TASK_ID --output-dir OUTPUT_DIR\nnode {baseDir}/scripts/piccc.mjs tasks --type image --status completed --page 1 --page-size 20\nnode {baseDir}/scripts/piccc.mjs auth login\nnode {baseDir}/scripts/piccc.mjs auth login --no-browser\nnode {baseDir}/scripts/piccc.mjs auth status\nnode {baseDir}/scripts/piccc.mjs auth logout\nnode {baseDir}/scripts/piccc.mjs credits\nnode {baseDir}/scripts/piccc.mjs --help"},{"language":"text","snippet":"安装技能 https://github.com/xiaoyaocz/piccc-ai-skills"},{"language":"bash","snippet":"npx skills add xiaoyaocz/piccc-ai-skills"},{"language":"text","snippet":"连接我的皮可AI账号"},{"language":"text","snippet":"用皮可AI生成一张 16:9 的电影感雪山日出图，完成后下载到 output/sunrise。"},{"language":"text","snippet":"用皮可AI把这段产品介绍生成中文旁白，先列出可用音色让我选择。"}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: piccc-ai\ndescription: Generate images, videos, and audio with Piccc AI (皮可AI), or query and download existing media tasks. Use when the user asks to create visual or audio media with Piccc AI, discover its current models or voices, check an asynchronous task, or download completed results. Do not use for editing existing local media.\n---\n\n# Piccc AI (皮可AI)\n\nUse `{baseDir}/scripts/piccc.mjs` for every Piccc AI request. `{baseDir}` is the directory containing this file.\n\n## Language\n\n- Reply in the user's current language.\n- Do not mix translated Chinese and English phrases in one response. The product name `Piccc AI (皮可AI)` is the only intentional bilingual form.\n- Treat CLI output as internal data and explain it naturally in the user's language.\n\n## Installation and first authorization\n\nWhen the skill has just been installed, tell the user in their current language that installation is complete, browser authorization is required, and the available credit balance will be reported afterward.\n\nBefore the first request, run `node {baseDir}/scripts/piccc.mjs auth status`. If it reports `authenticated: false`:\n\n1. Explain that the user must sign in and approve access in the browser.\n2. Run `node {baseDir}/scripts/piccc.mjs auth login` and wait for it to finish. The command opens the authorization page, receives the local callback, saves the API key, and queries the available balance. Do not ask the user to report completion.\n3. If the browser cannot be opened, share the displayed link and verification code while the command keeps waiting. Never ask the user to create, copy, or paste an API key.\n4. After authorization, read `available_credits` from the result and report the balance in the user's current language.\n5. If `available_credits` is unavailable, run `node {baseDir}/scripts/piccc.mjs credits` and then report the returned balance.\n\n## Workflow\n\n1. Unless the user names a model or asks for higher quality, run `models TYPE --economy` and use its selected model and defaults. Prefer lower cost over speed or quality.\n2. Let explicit user requirements override economy defaults. Otherwise use one output, the lowest supported resolution and quality, the shortest supported video duration, no audio or web search, and the lowest supported audio sample rate.\n3. If `special_offer` is true, warn the user in their current language before submission that special-offer models may be slower and less reliable.\n4. Select only values advertised by the live model response. Never invent model IDs, voice IDs, formats, ratios, resolutions, modes, or durations.\n5. Put long or multiline prompts in a UTF-8 file and use `--prompt-file`.\n6. When the user expects a finished file, use `--wait --output-dir DIR` in the initial request. Do not create another task while waiting.\n7. On timeout, return the task ID and latest status so the same task can be resumed. Treat `completed`, `failed`, and `cancelled` as terminal.\n8. For audio, call the initial 10 credits a precharge"},{"path":"README.md","content":"# Piccc AI（皮可AI）技能\n\n[![皮可AI，让创意一键成型](assets/piccc-ai-banner.png)](https://picccai.cn)\n\n让 Codex、OpenClaw 和其他兼容智能体技能标准的工具调用 Piccc AI（皮可AI），生成图片、视频和音频，并自动等待、下载生成结果。\n\n## 快速安装\n\n把下面这句话发给智能体：\n\n```text\n安装技能 https://github.com/xiaoyaocz/piccc-ai-skills\n```\n\n也可以使用 Skills 命令行工具：\n\n```bash\nnpx skills add xiaoyaocz/piccc-ai-skills\n```\n\n安装程序会识别仓库中的 `piccc-ai` 技能，并让你选择 Codex、OpenClaw 或其他已支持的智能体。\n\n## 连接皮可AI账号\n\n安装完成后，智能体应主动提示你完成账号授权。也可以直接告诉智能体：\n\n```text\n连接我的皮可AI账号\n```\n\n智能体会自动打开授权页面。登录 Piccc AI（皮可AI）并确认授权后，页面会跳转到本地完成连接，智能体会自动继续、查询当前可用积分并告诉你；API 密钥会创建并保存在本机，不需要复制密钥、配置环境变量或再回复“已经完成”。\n\n如果浏览器无法自动打开，命令仍会显示授权链接和验证码，手动打开链接即可。也可以使用 `node scripts/piccc.mjs auth login --no-browser` 明确启用原来的手动授权方式。\n\n## 使用\n\n安装后直接告诉智能体你要生成什么，例如：\n\n```text\n用皮可AI生成一张 16:9 的电影感雪山日出图，完成后下载到 output/sunrise。\n```\n\n```text\n用皮可AI把这段产品介绍生成中文旁白，先列出可用音色让我选择。\n```\n\n```text\n查询皮可AI任务 task_xxx，完成后把结果下载到 output/piccc。\n```\n\n技能会先读取后台当前可用的模型和参数，不会固定模型名称。未指定规格时默认优先选择更优惠的模型和低分辨率、低质量配置；使用特价模型前会提示生成可能较慢、不稳定。\n\n查看当前授权状态：\n\n```bash\nnode scripts/piccc.mjs auth status\n```\n\n取消本机授权：\n\n```bash\nnode scripts/piccc.mjs auth logout\n```\n\n## 手动安装\n\n下载仓库后，把根目录中的 `SKILL.md`、`agents`、`references` 和 `scripts` 复制到对应的 `piccc-ai` 目录：\n\n| 智能体 | 当前项目 | 全局使用 |\n| --- | --- | --- |\n| Codex | `.agents/skills/piccc-ai` | `~/.codex/skills/piccc-ai` |\n| OpenClaw | `/skills/piccc-ai` | `~/.openclaw/skills/piccc-ai` |\n\nOpenClaw 也可以直接安装本地目录：\n\n```bash\nopenclaw skills install ./piccc-ai-skills\n```\n\n脚本需要 Node.js 20 或更高版本，不需要安装 npm 依赖。\n\n## 本地检查\n\n```bash\nnpm run check\n```\n\n查看脚本支持的命令：\n\n```bash\nnode scripts/piccc.mjs --help\n```"},{"path":"_meta.json","content":"{\n \"ownerId\": \"kn797z5rnyaaady2chsrh164ws86gp7h\",\n \"slug\": \"piccc-ai-skills\",\n \"version\": \"0.1.4\",\n \"publishedAt\": 1785230989290\n}"},{"path":"references/api.md","content":"# Piccc AI 媒体接口\n\n## 连接账号\n\n首次使用时运行：\n\n```bash\nnode scripts/piccc.mjs auth login\n```\n\n命令会自动打开授权页面。登录 Piccc AI（皮可AI）并确认授权后，页面会跳转到本机回调页，脚本随后自动创建并保存 API 密钥、查询当前可用积分。命令会一直等待授权结果，不要让用户在对话中发送密钥，也不要要求用户另行回复“已完成”。浏览器无法打开时，使用命令显示的链接和验证码；`auth login --no-browser` 可保留纯手动流程。\n\n## 模型接口\n\n| 类型 | 接口 | 主要能力字段 |\n| --- | --- | --- |\n| 图片 | `GET /v1/images/models` | 比例、分辨率、质量、生成数量 |\n| 视频 | `GET /v1/videos/models` | 分辨率、比例、时长、生成模式、参考模式 |\n| 音频 | `GET /v1/audio/models` | 格式、采样率、参数范围、参考模式、音色 |\n\n每次创建任务前使用模型响应中的 `data[].id`。音色使用 `voice_presets[].id`，不要传显示名称。\n\n## 经济规格\n\n默认先运行：\n\n```bash\nnode scripts/piccc.mjs models image --economy\n```\n\n将 `image` 换成 `video` 或 `audio`。命令会返回推荐模型、`defaults`、`special_offer` 和 `warning`。\n\n- 用户没有指定时，使用 `defaults` 中的低分辨率、低质量、单个输出、最短视频时长和低采样率。\n- 默认关闭视频音频、联网搜索和不必要的参考素材。\n- 用户明确要求高清、高质量、更长时长或指定模型时，以用户要求为准。\n- `special_offer=true` 时，创建任务前提示：“特价模型可能生成较慢、稳定性较差。”\n- 即使没有运行 `--economy`，生成脚本也会为未传参数补上当前模型支持的经济规格。\n\n## 创建任务\n\n图片：`POST /v1/images/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`aspect_ratio`、`resolution`、`quality`、`n`、`web_search`、`external_id`\n\n视频：`POST /v1/videos/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`route_mode`、`reference_mode`、`resolution`、`aspect_ratio`、`duration_seconds`、`audio`、`web_search`、`external_id`\n\n音频：`POST /v1/audio/tasks`\n\n- 必填：`model`、`prompt`\n- 可选：`output_format`、`sample_rate`、`speech_rate`、`loudness_rate`、`pitch_rate`、`reference_mode`、`voice_id`、`audio_references`、`image_references`、`external_id`\n- `reference_mode=audio` 时传 `voice_id` 或 `audio_references`\n- `reference_mode=image` 时传 `image_references`\n- 提交时预扣 10 积分，完成后的 `actual_cost` 是最终费用\n\n## 查询任务\n\n- `GET /v1/tasks/{task_id}`：查询单个图片、视频或音频任务。\n- `GET /v1/tasks`：分页查询任务，可按 `type` 和 `status` 过滤。\n\n任务状态包括 `queued`、`running`、`completed`、`failed`、`cancelled`。只有 `completed` 状态下的 `outputs[].url` 才是最终结果。"},{"path":"skill-card.md","content":"## Description:\n\nGenerate images, videos, and audio with Piccc AI (皮可AI), and query or download completed media tasks.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[xiaoyaocz](https://clawhub.ai/user/xiaoyaocz)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nDevelopers and agent users use this skill to authorize a Piccc AI account, discover available media models and voices, submit image, video, or audio generation tasks, and wait for or download completed outputs.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: Linking a Piccc AI account and storing a local API key can expose account tokens on shared or untrusted machines.\n\nMitigation: Install only from a reviewed source, keep the credentials file private, and run auth logout when finished on shared machines.\n\nRisk: Custom Piccc AI API or authorization base URLs can redirect credential or task traffic outside the intended service.\n\nMitigation: Avoid setting PICCC_API_BASE_URL or PICCC_AUTH_BASE_URL outside controlled testing.\n\nRisk: Generated media downloads may place files in user-selected locations.\n\nMitigation: Review output directories before downloading completed image, video, or audio results.\n\n## Reference(s):\n\n- [Piccc AI media API reference](references/api.md)\n- [Piccc AI website](https://picccai.cn)\n- [ClawHub skill page](https://clawhub.ai/xiaoyaocz/skills/piccc-ai-skills)\n\n## Skill Output:\n\n**Output Type(s):** [text, markdown, shell commands, files, guidance]\n\n**Output Format:** [Markdown responses with shell commands and downloaded media files when tasks complete]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [May produce downloaded image, video, or audio files after asynchronous Piccc AI tasks complete.]\n\n## Skill Version(s):\n\n0.1.4 (source: server release evidence; package.json reports 0.1.0)\n\n## Ethical Considerations:\n\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment."}],"languages":[],"docsSourceLabel":"CLAWHUB","editorialOverview":null,"editorialQuality":{"score":100,"threshold":65,"status":"thin","wordCount":1719,"uniquenessScore":40,"reasons":["uniqueness-below-45"]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-10T22:52:22.565Z","emptyReason":"No screenshots, media assets, or demo links are available."},"primaryImageUrl":null,"mediaAssetCount":0,"assets":[],"demoUrl":null},"ownerResources":{"evidence":{"source":"unclaimed","verified":false,"confidence":"low","updatedAt":"2026-10-10T22:52:22.565Z","emptyReason":"This page has not been claimed by the agent owner."},"hasCustomPage":false,"customPageUpdatedAt":null,"customLinks":[],"structuredLinks":{"docsUrl":null,"demoUrl":null,"supportUrl":null,"pricingUrl":null,"statusUrl":null},"customPage":null},"relatedAgents":{"evidence":{"source":"protocol-neighbors","verified":false,"confidence":"medium","updatedAt":"2026-10-11T05:43:12.671Z","emptyReason":null},"items":[{"id":"8ebccd8e-3863-4187-8355-c3f14e1f9edf","entityType":"agent","canonicalPath":"/agent/iofficeai-aionui","slug":"iofficeai-aionui","name":"AionUi","description":"Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!","url":"https://github.com/iOfficeAI/AionUi","homepage":"https://www.aionui.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-10-09T19:11:12.944Z","createdAt":"2026-02-25T03:38:16.584Z","downloads":null},{"id":"b917f68a-ebff-438e-84f8-3f4b2494c0bc","entityType":"agent","canonicalPath":"/agent/activepieces-activepieces","slug":"activepieces-activepieces","name":"activepieces","description":"AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents","url":"https://github.com/activepieces/activepieces","homepage":"https://www.activepieces.com","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-15T02:22:12.426Z","createdAt":"2026-02-25T03:38:12.412Z","downloads":null},{"id":"5cb26759-3a39-483f-94cf-276a98c13bb8","entityType":"agent","canonicalPath":"/agent/cherryhq-cherry-studio","slug":"cherryhq-cherry-studio","name":"cherry-studio","description":"AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs","url":"https://github.com/CherryHQ/cherry-studio","homepage":"https://cherry-ai.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-11T14:38:40.986Z","createdAt":"2026-02-25T03:38:19.379Z","downloads":null},{"id":"6f6582d0-5d76-4f0f-b81d-86520247950b","entityType":"agent","canonicalPath":"/agent/copilotkit-copilotkit","slug":"copilotkit-copilotkit","name":"CopilotKit","description":"The Frontend for Agents & Generative UI. React + Angular","url":"https://github.com/CopilotKit/CopilotKit","homepage":"https://docs.copilotkit.ai","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-03-25T09:50:57.846Z","createdAt":"2026-02-25T03:39:14.617Z","downloads":null}],"links":{"hub":"/agent","source":"/agent/source/clawhub","protocols":[{"label":"OpenClaw","href":"/agent/protocol/openclew"}]}}}