{"id":"de0f779b-fff2-49d4-8480-c8bd19c2e255","entityType":"agent","slug":"clawhub-etmnb-bytedance-ai-image-gen","name":"Bytedance Ai Image Gen","canonicalUrl":"https://www.xpersona.co/agent/clawhub-etmnb-bytedance-ai-image-gen","canonicalPath":"/agent/clawhub-etmnb-bytedance-ai-image-gen","generatedAt":"2026-10-10T07:42:13.209Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"editorial-content","verified":true,"confidence":"high","updatedAt":"2026-10-10T04:06:12.435Z","emptyReason":null},"description":"ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。 支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。 参与火山协作奖励计划免费使用顶级生图模型。 智能调度模型，每模型每日18次，超限拒绝。 ⏱️ 生成需30-60秒，触发后请耐心等待。 Skill: Bytedance Ai Image Gen Owner: etmnb Summary: ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。 支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。 参与火山协作奖励计划免费使用顶级生图模型。 智能调度模型，每模型每日18次，超限拒绝。 ⏱️ 生成需30-60秒，触发后请耐心等待。 Tags: latest:2.4.2 Version history: v2.4.2 | 2026-05-28T03:04:09.927Z | auto bytedance-ai-image-gen v2.4.2 - No code or functionality changes. - Internal metadata updated only. - No impact on user experience or feature","descriptionLabel":"Technical summary","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 1.7K downloads reported by the source. Last updated 10/10/2026.","installCommand":"clawhub skill install s17d5t69ggjwkkgy7xx496krnx83pg6h:bytedance-ai-image-gen","sourceUrl":"https://clawhub.ai/etmnb/bytedance-ai-image-gen","homepage":"https://clawhub.ai/etmnb/skills/bytedance-ai-image-gen","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/etmnb/bytedance-ai-image-gen","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/etmnb/skills/bytedance-ai-image-gen","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":65,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。 支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。 参与火山协作奖励计划免费使用顶级生图模型。 智能调度模型，每模型每日18次，超限拒绝。 ⏱️ 生成需30-60秒，触发后请耐心等待。 Skill: Byt"},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-10T04:06:12.435Z","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":"2026-10-10T04:06:12.435Z","emptyReason":null},"stars":null,"forks":null,"downloads":1703,"packageName":null,"latestVersion":"2.4.2","tractionLabel":"1.7K downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-10T04:06:12.435Z","emptyReason":null},"lastUpdatedAt":"2026-10-10T04:06:12.435Z","lastCrawledAt":"2026-10-10T04:06:12.435Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-11T04:06:12.435Z","lastVerifiedAt":null,"highlights":[{"version":"2.4.2","createdAt":"2026-05-28T03:04:09.927Z","changelog":"bytedance-ai-image-gen v2.4.2 - No code or functionality changes. - Internal metadata updated only. - No impact on user experience or features.","fileCount":4,"zipByteSize":24779},{"version":"2.3.4","createdAt":"2026-05-28T03:02:54.324Z","changelog":"- Updated skill description and documentation for clearer usage instructions and parameter inference rules. - Added detailed step-by-step configuration guide, including required environment variables and console setup steps. - Emphasized strict operational guidelines: users must not be prompted for confirmation before execution, and all triggers must result in immediate action. - Clarified command formats for text-to-image, image-to-image, multi-image fusion, and usage synchronization. - Outlined daily model usage limits and automatic fallback behaviors. - Provided concise rules and size inference logic for streamlined skill activation and compliance.","fileCount":4,"zipByteSize":24664},{"version":"1.0.0","createdAt":"2026-05-28T02:32:16.181Z","changelog":"- Initial release of ByteDance AI Image Gen skill, supporting free image generation and editing using Doubao/ByteDance Seedream models. - Enables text-to-image, image-to-image, batch image creation, multi-image blending, and advanced image editing features. - Smart model routing with automatic fallback: Seedream-4.5 → 5.0-Lite → 4.0, each model limited to 18 uses per day. - CLI-based usage with auto parameter inference; image resolution and aspect ratio are detected or adapted automatically. - Simple setup with required API keys and model IDs; usage does not require user confirmation. - Fully documented commands, trigger patterns, and error handling.","fileCount":5,"zipByteSize":27821},{"version":"2.3.3","createdAt":"2026-05-27T18:38:46.329Z","changelog":"## bytedance-ai-image-gen 2.3.3 Changelog - Metadata file updated; version number remains at 2.3.0. - No user-facing changes or new features in this release.","fileCount":3,"zipByteSize":23371},{"version":"2.3.2","createdAt":"2026-05-27T18:29:30.389Z","changelog":"- Improved documentation with clearer setup instructions, usage commands, and strict behavior guidelines. - Enhanced parameter inference rules for image generation, editing, and fusion. - Clarified model selection process and limitations, including daily usage caps and feature support for each model. - Added detailed instructions for environment variable configuration and optional IAM integration. - Emphasized immediate command execution upon trigger, with strict rules to avoid unnecessary confirmations.","fileCount":3,"zipByteSize":23371},{"version":"2.3.1","createdAt":"2026-05-27T18:27:25.381Z","changelog":"- Update version from 2.3.0 to 2.3.1 in metadata. - No other content changes.","fileCount":3,"zipByteSize":23371},{"version":"2.4.0","createdAt":"2026-05-27T18:26:40.834Z","changelog":"**No major functional changes. Documentation has been simplified to make setup and usage clearer.** - Simplified and clarified initial setup instructions, including API Key、模型 ID 获取和 .env 配置流程。 - 优化行为规则，强调“触发即执行”“无需确认”。 - 用法示例和参数说明合并精简，去除重复内容。 - 摘要与描述更新，更突出参与火山协作奖励计划享受免费使用福利。 - 对常见场景、命令和模型说明做了聚合整理，方便查阅。","fileCount":4,"zipByteSize":24627},{"version":"1.1.0","createdAt":"2026-05-27T18:23:50.280Z","changelog":"- 改进文档结构，首次配置步骤更易懂，配置过程简化说明 - 丰富调用命令、参数示例，强化命令速查表 - 优化参数、模型、用法说明，明确各功能边界与限制 - 强调只需首次配置，已配好环境可直接用命令，无需重复配置 - 保持现有功能与触发模式不变，主要为文档内容精简与指导性提升","fileCount":3,"zipByteSize":23371}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s17d5t69ggjwkkgy7xx496krnx83pg6h:bytedance-ai-image-gen","setupComplexity":"low","setupSteps":["Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.","Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data."],"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-etmnb-bytedance-ai-image-gen/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-etmnb-bytedance-ai-image-gen/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-etmnb-bytedance-ai-image-gen/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-etmnb-bytedance-ai-image-gen/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-etmnb-bytedance-ai-image-gen/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-etmnb-bytedance-ai-image-gen/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-10T07:42:13.207Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-etmnb-bytedance-ai-image-gen/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-etmnb-bytedance-ai-image-gen/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-etmnb-bytedance-ai-image-gen/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-etmnb-bytedance-ai-image-gen/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":"high","updatedAt":"2026-10-10T04:06:12.435Z","emptyReason":null},"readme":"Skill: Bytedance Ai Image Gen\n\nOwner: etmnb\n\nSummary: ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。 支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。 参与火山协作奖励计划免费使用顶级生图模型。 智能调度模型，每模型每日18次，超限拒绝。 ⏱️ 生成需30-60秒，触发后请耐心等待。\n\nTags: latest:2.4.2\n\nVersion history:\n\nv2.4.2 | 2026-05-28T03:04:09.927Z | auto\n\nbytedance-ai-image-gen v2.4.2\n\n- No code or functionality changes.  \n- Internal metadata updated only.  \n- No impact on user experience or features.\n\nv2.3.4 | 2026-05-28T03:02:54.324Z | auto\n\n- Updated skill description and documentation for clearer usage instructions and parameter inference rules.\n- Added detailed step-by-step configuration guide, including required environment variables and console setup steps.\n- Emphasized strict operational guidelines: users must not be prompted for confirmation before execution, and all triggers must result in immediate action.\n- Clarified command formats for text-to-image, image-to-image, multi-image fusion, and usage synchronization.\n- Outlined daily model usage limits and automatic fallback behaviors.\n- Provided concise rules and size inference logic for streamlined skill activation and compliance.\n\nv1.0.0 | 2026-05-28T02:32:16.181Z | auto\n\n- Initial release of ByteDance AI Image Gen skill, supporting free image generation and editing using Doubao/ByteDance Seedream models.\n- Enables text-to-image, image-to-image, batch image creation, multi-image blending, and advanced image editing features.\n- Smart model routing with automatic fallback: Seedream-4.5 → 5.0-Lite → 4.0, each model limited to 18 uses per day.\n- CLI-based usage with auto parameter inference; image resolution and aspect ratio are detected or adapted automatically.\n- Simple setup with required API keys and model IDs; usage does not require user confirmation.\n- Fully documented commands, trigger patterns, and error handling.\n\nv2.3.3 | 2026-05-27T18:38:46.329Z | auto\n\n## bytedance-ai-image-gen 2.3.3 Changelog\n\n- Metadata file updated; version number remains at 2.3.0.\n- No user-facing changes or new features in this release.\n\nv2.3.2 | 2026-05-27T18:29:30.389Z | auto\n\n- Improved documentation with clearer setup instructions, usage commands, and strict behavior guidelines.\n- Enhanced parameter inference rules for image generation, editing, and fusion.\n- Clarified model selection process and limitations, including daily usage caps and feature support for each model.\n- Added detailed instructions for environment variable configuration and optional IAM integration.\n- Emphasized immediate command execution upon trigger, with strict rules to avoid unnecessary confirmations.\n\nv2.3.1 | 2026-05-27T18:27:25.381Z | auto\n\n- Update version from 2.3.0 to 2.3.1 in metadata.\n- No other content changes.\n\nv2.4.0 | 2026-05-27T18:26:40.834Z | auto\n\n**No major functional changes. Documentation has been simplified to make setup and usage clearer.**\n\n- Simplified and clarified initial setup instructions, including API Key、模型 ID 获取和 .env 配置流程。\n- 优化行为规则，强调“触发即执行”“无需确认”。\n- 用法示例和参数说明合并精简，去除重复内容。\n- 摘要与描述更新，更突出参与火山协作奖励计划享受免费使用福利。\n- 对常见场景、命令和模型说明做了聚合整理，方便查阅。\n\nv1.1.0 | 2026-05-27T18:23:50.280Z | auto\n\n- 改进文档结构，首次配置步骤更易懂，配置过程简化说明\n- 丰富调用命令、参数示例，强化命令速查表\n- 优化参数、模型、用法说明，明确各功能边界与限制\n- 强调只需首次配置，已配好环境可直接用命令，无需重复配置\n- 保持现有功能与触发模式不变，主要为文档内容精简与指导性提升\n\nv1.0.9 | 2026-05-27T18:21:47.799Z | auto\n\n- 文档内容精简与结构优化，提升配置与使用方法的清晰度\n- 配置指引独立为“首次配置”，强调只需一次配置即可直接用\n- 明确调用方式与行为规则，突出无需重复确认、即触发即执行\n- 更新参数推断与典型命令说明，便于用户快速上手\n- 触发词和限制说明保持与旧版一致\n\nv2.3.0 | 2026-05-25T17:03:06.002Z | user\n\nv2.3.0: 模型能力矩阵+已知限制文档化, 组图自动补前缀, 多图融合payload限制说明\n\nv2.2.9 | 2026-05-25T16:52:50.231Z | user\n\nv2.2.9: 修复组图只出1张Bug(自动补生成N张前缀), 3图融合验证通过, 统一版本号\n\nv2.2.8 | 2026-05-25T16:33:02.728Z | user\n\nv2.2.8: 修复文生组图/图生组图只出1张的Bug(API要求prompt明确写生成N张，脚本自动补前缀)\n\nv2.2.7 | 2026-05-25T15:21:52.678Z | auto\n\n**Summary:**\nIntroduces debugging and testing utilities to enhance troubleshooting and validation of the image generation process.\n\n- Added multiple debug scripts and test utilities (e.g., debug_history.py, debug_make_filename.py, verify_prompts.py) for troubleshooting and validating filename creation, history tracking, and prompt verification.\n- Introduced image_gen_history.json to support detailed logging of image generation history.\n- No changes to core functionality or user-facing commands.\n- SKILL.md and metadata updated only for version/control purposes; main behavior and documentation remain unchanged.\n\nv2.2.6 | 2026-05-25T15:15:06.043Z | user\n\nv2.2.6: 修复CDN下载失败(直连被墙->自动回退代理), 修复URL截断和历史记录存储, 修复_用法示例输出完整URL\n\nv2.2.5 | 2026-05-25T15:06:27.284Z | user\n\nv2.2.5: 调用方式增加输出规范，要求 Agent 必须报告本地文件路径而非仅 URL\n\nv2.2.4 | 2026-05-25T15:03:50.389Z | user\n\nv2.2.4: 新增'调用方式'命令模板，Agent 触发后直接执行命令而非空回复\n\nv2.2.3 | 2026-05-25T13:11:44.709Z | user\n\nv2.2.3: 工作流程中移除 API 端点 URL，改为通用描述\n\nv2.2.2 | 2026-05-25T13:06:19.358Z | user\n\nv2.2.2: 移除无效命令、修正 metadata 安全声明\n\nv2.2.1 | 2026-05-23T13:41:20.592Z | auto\n\n- 文档命令示例和参数说明简化，使用“命令参考”替代具体 CLI 示例表格。\n- 去除部分详细用法和用户输入→CLI映射，推荐通过 `--help` 参数查看命令用法。\n- 针对命令和文档进行了结构优化，主逻辑、触发规则和配置说明未变。\n- 无功能代码变动，仅文档（SKILL.md）简化和整理。\n\nv2.1.9 | 2026-05-22T15:08:48.269Z | auto\n\nbytedance-ai-image-gen 2.1.9\n\n- Removed openclaw.json for simplification.\n- Updated SKILL.md: clarified \"禁止确认配额消耗\"与\"禁止确认执行\"等行为规则条目表述，显式说明执行流程无需用户二次确认。\n- 规则说明部分重写和精简，无改动产品功能。\n- 各功能、常见错误等文档描述同步小幅调整至最新版。\n- 不涉及实际脚本/功能变更，仅文档修正及移除冗余文件。\n\nv2.1.8 | 2026-05-22T09:01:06.453Z | auto\n\nbytedance-ai-image-gen 2.1.8\n\n- Improved and updated documentation for CLI-first usage, adding clear CLI examples for all features.\n- Clarified that CLI commands are now recommended as the default interaction mode, with detailed command reference and invocation patterns.\n- Enhanced parameter explanation and mapping between common user requests and CLI parameters.\n- Minor corrections to file and directory naming conventions in documentation (e.g., `Temp/`).\n- Internal refactoring in `doubao_image_gen.py` to fully support and synchronize with updated CLI usage and argument parsing.\n- No breaking changes to core image generation functionality.\n\nv2.1.6 | 2026-05-22T07:12:41.158Z | auto\n\n**bytedance-ai-image-gen 2.1.6**\n\n- Added new trigger_patterns for richer phrase and command coverage.\n- Added a summary and multiple tags to improve Skill discoverability.\n- Clarified .env required environment variable descriptions; now consistent with API requirements.\n- Updated documentation to support user-specified model selection using friendly abbreviations (e.g., \"4.0\", \"5l\").\n- Watermark control details added: switchable per .env or single command.\n- Refined error and FAQ documentation for clearer troubleshooting and usage guidance.\n\nv2.1.0 | 2026-05-22T05:11:49.111Z | user\n\nv2.1.0: CLI新增seq/fusion命令支持组图/融合, 5模式全通过测试, 触发词移至SKILL.md\n\nv2.0.5 | 2026-05-22T04:39:43.952Z | user\n\nRemoved trigger detection from Python (now in SKILL.md), pure SDK\n\nv2.0.4 | 2026-05-22T04:30:12.921Z | user\n\nEnhanced trigger words, added features listing, 30-60s wait notice, 图生组图 verified\n\nv2.0.3 | 2026-05-21T12:41:57.553Z | auto\n\n**New: Supports direct usage syncing with Volcengine console via IAM/endpoint.**\n\n- 支持填写 Volcengine 控制台端点 ID (`DOUBAO_ENDPOINT_*_ID`) 实现同步控制台用量（免消耗配额）。\n- 新增 python doubao_image_gen.py sync 命令，同步并查询控制台用量。\n- SKILL.md 增加 IAM 密钥和端点 ID 配置说明。\n- 版本号更新为 2.0.3。\n\nv2.0.2 | 2026-05-21T06:47:52.779Z | auto\n\n- Improved documentation for skill usage, adding more explicit Python call examples for different image generation scenarios.\n- Clarified and expanded calling demonstrations for single image, batch image, image-to-image, and multi-image fusion cases.\n- No changes to APIs or skill logic; update is documentation-only.\n\nv2.0.1 | 2026-05-21T06:35:48.324Z | auto\n\n**Changelog for bytedance-ai-image-gen 2.0.1**\n\n- 默认画质由“2K”升级为“3K”，API 调用缺省不再传比例字段，由后端自适应分辨率。\n- 更新了尺寸推断与参数传递规则，确保只有用户明确指定时才传 size/ratio，否则自动适配。\n- 文档同步更新，所有示例和表格默认以“3K”为缺省画质说明。\n- 细化了尺寸/比例推断表述，提升调用体验和输出一致性。\n\nv2.0.0 | 2026-05-21T05:37:37.495Z | auto\n\nbytedance-ai-image-gen 2.0.0\n\n- 增强分辨率自适应：支持模型自定宽高比，预设 2K/3K/4K 不再等同于固定尺寸，像素依次 ≤ 4.2M / 9.4M / 16.8M。\n- 文档更新：SKILL.md 描述和参数表同步本地变化，澄清分辨率逻辑、“自适应”规则等细节。\n- 保持原有调用、行为规则与模型日用量限制。\n\nv1.1.9 | 2026-05-21T05:14:46.645Z | auto\n\n- 更新版本号为 1.1.9（原为 1.1.8）\n- 仅文档（SKILL.md）内容更新，功能未做更改\n- 说明、规则、配置说明及使用文档内容无实质性调整，仅反映版本号更新\n\nv1.1.8 | 2026-05-21T05:05:07.230Z | auto\n\nVersion 1.1.8\n\n- Updated documentation in SKILL.md for clarity and conciseness.\n- No code or behavior changes; this is a documentation-only update.\n- API usage instructions and behavioral rules remain unchanged.\n\nv1.1.6 | 2026-05-21T04:58:02.680Z | auto\n\n**bytedance-ai-image-gen v1.1.6 – Changelog**\n\n- 文档链接更新为官方 1541523 文档页，保障中英文一致和内容准确。\n- SKILL.md version 字段更新，描述内容同步小幅调整以配合官网链接变化。\n- 无接口或API功能代码更改。\n\nv1.1.3 | 2026-05-21T04:47:08.618Z | auto\n\n- Update to version 1.1.3.\n- Adjusted description: removed \"新用户通过火山「协作奖励计划」免费使用…\" → now simply \"用户通过火山「协作奖励计划」免费使用…\"\n- Minor clarification in the \"协作奖励计划\" section: removed \"新用户注册\" wording.\n- Documentation improvements in SKILL.md, including added links to API documentation in both Chinese and English.\n- No changes to required environment variables or installation instructions.\n\nv1.1.2 | 2026-05-21T04:26:54.541Z | auto\n\n- Homepage/doc links updated to new Volcengine documentation URLs in SKILL.md and metadata.\n- No functional changes; instructions and API usage remain the same.\n\nv1.0.8 | 2026-05-21T04:08:05.289Z | user\n\nCLI: support --size/--ratio named args + generate alias\n\nv1.0.7 | 2026-05-21T03:09:45.690Z | auto\n\n**Summary:**  \nimage_gen_history.json has been removed and related logic updated; added support for temp/ local cache directory; general improvements and metadata/version updates.\n\n- Removed image_gen_history.json and updated history/recording logic accordingly.\n- Added/updated references to temp/ directory for local cache, with automatic creation/cleanup.\n- Updated SKILL.md with new information about file structure and support/feedback section.\n- Bumped version to 1.0.7 and refreshed metadata in _meta.json and openclaw.json.\n- General improvements to documentation and internal handling.\n\nv1.0.5 | 2026-05-21T01:55:24.816Z | auto\n\nVersion 1.0.5\n\n- Added a section introducing the \"火山协作奖励计划\" (Volcengine Collaboration Reward Program), highlighting free access to top models for new users.\n- No changes to code or functionality; documentation (SKILL.md) only updated.\n\nv1.0.4 | 2026-05-21T01:52:47.296Z | auto\n\n- 新增：说明新用户可通过火山「协作奖励计划」免费使用顶级模型（SKILL.md）。\n- 优化：行为规则描述更简洁，突出自动推断分辨率、用量自动限制。\n- 补充：明确每日每模型18次用量限制，超限直接拒绝调用。\n- 文档微调：简化描述和用例，提升可读性。\n- 版本号更新为 1.0.4。\n\nv1.0.3 | 2026-05-21T01:39:06.987Z | auto\n\n- Improved automatic size and aspect ratio inference: “2K” and “4K” now use adaptive (default 16:9) proportions instead of always square.\n- Added new “3K” preset (3072²), used with 16:9/9:16 triggers for better widescreen/vertical results.\n- Updated handling for specific prompt words: e.g., “宽屏/横版/16:9” triggers 3K+16:9, “竖版/9:16” triggers 3K+9:16, “电影” triggers 4K+21:9.\n- Documentation and usage instructions updated to reflect the new dimension rules and selection logic.\n\nv1.0.1 | 2026-05-21T01:17:27.413Z | auto\n\nByteDance AI Image Gen v1.0.1\n\n- Initial release of the skill, providing image generation via Doubao/ByteDance Seedream models.\n- Supports text-to-image, image-to-image (local path, URL, base64), multi-image blending, and sequential image generation.\n- Intelligent scheduling and daily usage limit for each model (Seedream-4.5, 5.0 Lite, 4.0; 18/day each).\n- Triggered only by specific keywords and phrases related to image generation or editing.\n- Strict rules: never confirm usage or parameters with the user; automatically infer resolution/ratio.\n- Includes CLI usage instructions, environment variable setup, and a robust workflow for image creation and management.\n\nArchive index:\n\nArchive v2.4.2: 4 files, 24779 bytes\n\nFiles: _meta.json (141b), doubao_image_gen.py (71543b), skill-card.md (2627b), SKILL.md (6839b)\n\nFile v2.4.2:SKILL.md\n\n---\nname: bytedance-ai-image-gen\ndescription: >\n  ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。\n  支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。\n  参与火山协作奖励计划免费使用顶级生图模型。\n  智能调度模型，每模型每日18次，超限拒绝。\n  ⏱️ 生成需30-60秒，触发后请耐心等待。\nsummary: \"ByteDance AI Image Gen — 调用豆包 Seedream 生成/编辑图片，支持文生图、图生图、组图、多图融合，参与火山协作奖励计划免费使用\"\ntags:\n  image-generation: \"2.1.0\"\n  byteimage: \"2.1.0\"\n  seedream: \"2.1.0\"\n  text-to-image: \"2.1.0\"\n  image-to-image: \"2.1.0\"\n  volcengine: \"2.1.0\"\nversion: 2.3.0\ntrigger_patterns:\n  - \"生成图片\"\n  - \"画一张图\"\n  - \"创建图片\"\n  - \"制作图片\"\n  - \"画图\"\n  - \"帮我画\"\n  - \"帮我生成\"\n  - \"我想生成\"\n  - \"搞个图\"\n  - \"来张图\"\n  - \"做张图\"\n  - \"generate image\"\n  - \"create image\"\n  - \"txt2img\"\n  - \"please generate\"\n  - \"please draw\"\n  - \"图生图\"\n  - \"图片编辑\"\n  - \"编辑图片\"\n  - \"改图\"\n  - \"换背景\"\n  - \"换风格\"\n  - \"img2img\"\n  - \"image to image\"\n  - \"edit this image\"\n  - \"edit image\"\n  - \"融合\"\n  - \"合成\"\n  - \"两张图片\"\n  - \"多张图片\"\n  - \"拼合\"\n  - \"合图\"\n  - \"blend\"\n  - \"fuse\"\n  - \"merge\"\n  - \"combine\"\n  - \"文生图\"\n  - \"生图\"\n  - \"图片生成\"\n  - \"图像生成\"\n  - \"生成图像\"\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - python\n      env:\n        - ARK_API_KEY\n        - DOUBAO_MODEL_45_ID\n        - DOUBAO_MODEL_50L_ID\n        - DOUBAO_MODEL_40_ID\n    primaryEnv: ARK_API_KEY\n    envVars:\n      - name: ARK_API_KEY\n        required: true\n        description: 火山方舟 API Key，在 console.volcengine.com/ark → API Key 管理 创建\n      - name: DOUBAO_MODEL_45_ID\n        required: true\n        description: Seedream-4.5 接入点 ID，在 console.volcengine.com/ark → 在线推理 获取\n      - name: DOUBAO_MODEL_50L_ID\n        required: true\n        description: Seedream-5.0-Lite 接入点 ID\n      - name: DOUBAO_MODEL_40_ID\n        required: true\n        description: Seedream-4.0 接入点 ID\n      - name: VOLCENGINE_ACCESS_KEY\n        required: false\n        description: 火山 IAM Access Key，在 console.volcengine.com/iam 创建，用于自动同步用量\n      - name: VOLCENGINE_SECRET_KEY\n        required: false\n        description: 火山 IAM Secret Key，与 ACCESS_KEY 配套\n    emoji: \"🎨\"\n    homepage: https://www.volcengine.com/docs/82379/1541523\n    install:\n      - kind: uv\n        package: requests\n---\n\n# ByteDance AI Image Gen — 豆包 Seedream 图片生成\n\n调用火山方舟 Seedream [API 文档](https://www.volcengine.com/docs/82379/1541523?lang=zh) 生成/编辑图片，自动选择模型并限制用量。参与火山协作奖励计划免费使用顶级生图模型。\n\n> ⚠️ **配置只需一次！** 如果 `.env` 文件已存在且包含 `ARK_API_KEY` 和 3 个模型 ID，说明已配置过，**直接跳到\"调用方式\"执行命令，不要重新配置。**\n\n## 🚀 首次配置\n\n### 1. 获取 API Key\n\n1. 打开 https://console.volcengine.com/ark 注册/登录\n2. 左侧菜单 → **API Key 管理** → 创建 API Key → 复制保存\n\n### 2. 创建模型接入点\n\n在同一个控制台，左侧菜单 → **在线推理** → 创建推理接入点，选以下 3 个模型：\n\n| 环境变量名 | 模型 | 说明 |\n|-----------|------|------|\n| `DOUBAO_MODEL_45_ID` | Seedream-4.5 | 主力模型，所有模式优先 |\n| `DOUBAO_MODEL_50L_ID` | Seedream-5.0-Lite | 轻量备选 |\n| `DOUBAO_MODEL_40_ID` | Seedream-4.0 | 仅支持文生图 |\n\n每个接入点创建后会得到一个 `ep-xxxxx` 格式的 ID，复制保存。\n\n### 3. 配置 .env 文件\n\n在 Skill 目录下创建 `.env` 文件，填入你的 Key 和接入点 ID：\n\n```bash\nARK_API_KEY=你的API Key\nDOUBAO_MODEL_45_ID=ep-xxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxx\n```\n\n> 可选：加上 IAM 密钥可自动同步控制台用量\n> ```bash\n> VOLCENGINE_ACCESS_KEY=你的Access Key\n> VOLCENGINE_SECRET_KEY=你的Secret Key\n> ```\n\n### 4. 测试\n\n```bash\npython doubao_image_gen.py --help\n```\n\n不报错，就配置好了。\n\n---\n\n## ⚡ 调用方式 — 触发后必须立即执行命令\n\n> **这是操作手册，不是参考文档。检测到触发词后，直接复制对应命令执行，禁止只回复文字。**\n\n确认 Skill 目录后，先 `cd` 进去，再执行对应命令。Skill 安装在: `.openclaw/workspace/skills/bytedance-ai-image-gen/`\n\n执行成功后，脚本会输出 `✅ 生成成功!` 及本地文件路径。**你必须把文件路径告知用户**。\n\n⚠️ **严禁**: 只回复\"正在生成...\"而不执行命令。必须跑 `python doubao_image_gen.py ...`。\n\n### 命令\n\n| 命令 | 用途 | 示例 |\n|------|------|------|\n| `gen \"描述\" [尺寸] [--ratio 比例]` | 文生图 | `gen \"洪崖洞夜景\" 4K --ratio 16:9` |\n| `img2img \"描述\" \"图片\" [尺寸]` | 图生图 | `img2img \"换成夜景\" \"photo.jpg\" 4K` |\n| `seq \"描述\" [--num 数量]` | 组图 | `seq \"猫咪\" --num 3` |\n| `seq \"描述\" --refs img1 img2` | 多图融合 | `seq \"融合\" --refs a.jpg b.jpg` |\n| `sync` | 同步控制台用量 | |\n\n### 参数\n\n| 参数 | 说明 |\n|------|------|\n| `尺寸` | `3K`(默认) / `4K`(高清) / 自定义 |\n| `--ratio` | `16:9`(宽屏) / `9:16`(竖屏) / `21:9`(电影) |\n| `--no-watermark` | 去水印 |\n| `--num` | 组图数量 |\n\n### 尺寸推断规则\n\n- 没提画质 → `3K`\n- 说\"4K/高清/超清\" → `4K`\n- 说\"16:9/宽屏\" → `3K --ratio 16:9`\n- 说\"9:16/竖屏\" → `3K --ratio 9:16`\n- 说\"21:9/电影\" → `4K --ratio 21:9`\n- 图生图 → 自动检测参考图分辨率\n\n### 模型选择\n\n- 自动降级: 4.5 → 5.0Lite → 4.0，每模型日限 18 次\n- 用户可指定简称: `4.5` / `5l` / `4.0`\n- 4.0 仅支持文生图，其他模式自动降级\n\n### 水印\n\n- 默认带水印，`.env` 设 `DEFAULT_WATERMARK=false` 永久关闭\n- 单次覆盖: `--no-watermark`\n\n---\n\n## 🚫 行为规则（最高优先级 — 必须遵守！）\n\n### 1. 禁止确认配额消耗\n- ✅ 用户调用 Skill 即表示知情并同意消耗配额，**直接执行，一个字都别问**\n\n### 2. 禁止确认执行\n- ✅ 触发即执行，不要等用户二次确认\n\n**总结：触发 → 推断参数 → 直接执行。中间不要有任何回合的确认。**\n\n### 3. 参数推断规则\n- 用户说\"生成/画/创建\"图片 → `gen`\n- 用户说\"编辑/改/换风格\"图片 → `img2img`\n- 用户说\"融合/合成\"多图 → `seq --refs`\n- 用户有额外要求 → 加到描述里\n\n---\n\n## 限制\n\n- 每个模型每天 18 次，超了自动换下一个\n- 4.0 仅支持文生图\n- 多图融合参考图 base64 后不能超过 ~15MB\n- 组图上限 15 张\n\nFile v2.4.2:_meta.json\n\n{\n  \"ownerId\": \"kn75s192k98dtwr83bft4myd0582ytyg\",\n  \"slug\": \"bytedance-ai-image-gen\",\n  \"version\": \"2.4.2\",\n  \"publishedAt\": 1779937449927\n}\n\nFile v2.4.2:skill-card.md\n\n## Description:\n\nByteDance AI Image Gen calls Volcengine Ark Doubao Seedream models to generate and edit single images, image sets, image-to-image edits, and multi-image fusions with automatic model selection and daily usage limits.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[etmnb](https://clawhub.ai/user/etmnb)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nDevelopers and external users use this skill to generate or edit images with Volcengine Ark Doubao Seedream models from natural-language prompts, reference images, or multiple images. It helps agents choose model and size parameters, run the Python CLI, and return generated local image paths.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: Prompts and reference images may be sent to Volcengine services.\n\nMitigation: Use only prompts and images that are acceptable to share with Volcengine; avoid sensitive local images.\n\nRisk: The skill can store prompts, image URLs, usage counts, generated files, and temporary copies locally.\n\nMitigation: Review and periodically clear the local history, Image, and Temp directories, and set storage paths deliberately when needed.\n\nRisk: The skill instructs agents to execute generation without a second confirmation, which can consume service quota.\n\nMitigation: Review trigger behavior before installation and restrict use to contexts where automatic execution is acceptable.\n\nRisk: API keys and model endpoint identifiers are required for operation.\n\nMitigation: Store secrets in secure environment management rather than a skill-directory .env file when possible.\n\n## Reference(s):\n\n- [ClawHub Skill Page](https://clawhub.ai/etmnb/skills/bytedance-ai-image-gen)\n- [Volcengine Ark Seedream API Documentation](https://www.volcengine.com/docs/82379/1541523)\n- [Volcengine Ark Console](https://console.volcengine.com/ark)\n\n## Skill Output:\n\n**Output Type(s):** [Shell commands, Files, Guidance]\n\n**Output Format:** [Console text with generated image file paths and status details]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [May save generated images under a local Image directory and keep seven-day prompt and usage history.]\n\n## Skill Version(s):\n\n2.4.2 (source: server release metadata; artifact frontmatter reports 2.3.0 and artifact _meta.json reports 2.3.1)\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\nArchive v2.3.4: 4 files, 24664 bytes\n\nFiles: _meta.json (141b), doubao_image_gen.py (71543b), skill-card.md (2478b), SKILL.md (6839b)\n\nFile v2.3.4:SKILL.md\n\n---\nname: bytedance-ai-image-gen\ndescription: >\n  ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。\n  支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。\n  参与火山协作奖励计划免费使用顶级生图模型。\n  智能调度模型，每模型每日18次，超限拒绝。\n  ⏱️ 生成需30-60秒，触发后请耐心等待。\nsummary: \"ByteDance AI Image Gen — 调用豆包 Seedream 生成/编辑图片，支持文生图、图生图、组图、多图融合，参与火山协作奖励计划免费使用\"\ntags:\n  image-generation: \"2.1.0\"\n  byteimage: \"2.1.0\"\n  seedream: \"2.1.0\"\n  text-to-image: \"2.1.0\"\n  image-to-image: \"2.1.0\"\n  volcengine: \"2.1.0\"\nversion: 2.3.0\ntrigger_patterns:\n  - \"生成图片\"\n  - \"画一张图\"\n  - \"创建图片\"\n  - \"制作图片\"\n  - \"画图\"\n  - \"帮我画\"\n  - \"帮我生成\"\n  - \"我想生成\"\n  - \"搞个图\"\n  - \"来张图\"\n  - \"做张图\"\n  - \"generate image\"\n  - \"create image\"\n  - \"txt2img\"\n  - \"please generate\"\n  - \"please draw\"\n  - \"图生图\"\n  - \"图片编辑\"\n  - \"编辑图片\"\n  - \"改图\"\n  - \"换背景\"\n  - \"换风格\"\n  - \"img2img\"\n  - \"image to image\"\n  - \"edit this image\"\n  - \"edit image\"\n  - \"融合\"\n  - \"合成\"\n  - \"两张图片\"\n  - \"多张图片\"\n  - \"拼合\"\n  - \"合图\"\n  - \"blend\"\n  - \"fuse\"\n  - \"merge\"\n  - \"combine\"\n  - \"文生图\"\n  - \"生图\"\n  - \"图片生成\"\n  - \"图像生成\"\n  - \"生成图像\"\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - python\n      env:\n        - ARK_API_KEY\n        - DOUBAO_MODEL_45_ID\n        - DOUBAO_MODEL_50L_ID\n        - DOUBAO_MODEL_40_ID\n    primaryEnv: ARK_API_KEY\n    envVars:\n      - name: ARK_API_KEY\n        required: true\n        description: 火山方舟 API Key，在 console.volcengine.com/ark → API Key 管理 创建\n      - name: DOUBAO_MODEL_45_ID\n        required: true\n        description: Seedream-4.5 接入点 ID，在 console.volcengine.com/ark → 在线推理 获取\n      - name: DOUBAO_MODEL_50L_ID\n        required: true\n        description: Seedream-5.0-Lite 接入点 ID\n      - name: DOUBAO_MODEL_40_ID\n        required: true\n        description: Seedream-4.0 接入点 ID\n      - name: VOLCENGINE_ACCESS_KEY\n        required: false\n        description: 火山 IAM Access Key，在 console.volcengine.com/iam 创建，用于自动同步用量\n      - name: VOLCENGINE_SECRET_KEY\n        required: false\n        description: 火山 IAM Secret Key，与 ACCESS_KEY 配套\n    emoji: \"🎨\"\n    homepage: https://www.volcengine.com/docs/82379/1541523\n    install:\n      - kind: uv\n        package: requests\n---\n\n# ByteDance AI Image Gen — 豆包 Seedream 图片生成\n\n调用火山方舟 Seedream [API 文档](https://www.volcengine.com/docs/82379/1541523?lang=zh) 生成/编辑图片，自动选择模型并限制用量。参与火山协作奖励计划免费使用顶级生图模型。\n\n> ⚠️ **配置只需一次！** 如果 `.env` 文件已存在且包含 `ARK_API_KEY` 和 3 个模型 ID，说明已配置过，**直接跳到\"调用方式\"执行命令，不要重新配置。**\n\n## 🚀 首次配置\n\n### 1. 获取 API Key\n\n1. 打开 https://console.volcengine.com/ark 注册/登录\n2. 左侧菜单 → **API Key 管理** → 创建 API Key → 复制保存\n\n### 2. 创建模型接入点\n\n在同一个控制台，左侧菜单 → **在线推理** → 创建推理接入点，选以下 3 个模型：\n\n| 环境变量名 | 模型 | 说明 |\n|-----------|------|------|\n| `DOUBAO_MODEL_45_ID` | Seedream-4.5 | 主力模型，所有模式优先 |\n| `DOUBAO_MODEL_50L_ID` | Seedream-5.0-Lite | 轻量备选 |\n| `DOUBAO_MODEL_40_ID` | Seedream-4.0 | 仅支持文生图 |\n\n每个接入点创建后会得到一个 `ep-xxxxx` 格式的 ID，复制保存。\n\n### 3. 配置 .env 文件\n\n在 Skill 目录下创建 `.env` 文件，填入你的 Key 和接入点 ID：\n\n```bash\nARK_API_KEY=你的API Key\nDOUBAO_MODEL_45_ID=ep-xxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxx\n```\n\n> 可选：加上 IAM 密钥可自动同步控制台用量\n> ```bash\n> VOLCENGINE_ACCESS_KEY=你的Access Key\n> VOLCENGINE_SECRET_KEY=你的Secret Key\n> ```\n\n### 4. 测试\n\n```bash\npython doubao_image_gen.py --help\n```\n\n不报错，就配置好了。\n\n---\n\n## ⚡ 调用方式 — 触发后必须立即执行命令\n\n> **这是操作手册，不是参考文档。检测到触发词后，直接复制对应命令执行，禁止只回复文字。**\n\n确认 Skill 目录后，先 `cd` 进去，再执行对应命令。Skill 安装在: `.openclaw/workspace/skills/bytedance-ai-image-gen/`\n\n执行成功后，脚本会输出 `✅ 生成成功!` 及本地文件路径。**你必须把文件路径告知用户**。\n\n⚠️ **严禁**: 只回复\"正在生成...\"而不执行命令。必须跑 `python doubao_image_gen.py ...`。\n\n### 命令\n\n| 命令 | 用途 | 示例 |\n|------|------|------|\n| `gen \"描述\" [尺寸] [--ratio 比例]` | 文生图 | `gen \"洪崖洞夜景\" 4K --ratio 16:9` |\n| `img2img \"描述\" \"图片\" [尺寸]` | 图生图 | `img2img \"换成夜景\" \"photo.jpg\" 4K` |\n| `seq \"描述\" [--num 数量]` | 组图 | `seq \"猫咪\" --num 3` |\n| `seq \"描述\" --refs img1 img2` | 多图融合 | `seq \"融合\" --refs a.jpg b.jpg` |\n| `sync` | 同步控制台用量 | |\n\n### 参数\n\n| 参数 | 说明 |\n|------|------|\n| `尺寸` | `3K`(默认) / `4K`(高清) / 自定义 |\n| `--ratio` | `16:9`(宽屏) / `9:16`(竖屏) / `21:9`(电影) |\n| `--no-watermark` | 去水印 |\n| `--num` | 组图数量 |\n\n### 尺寸推断规则\n\n- 没提画质 → `3K`\n- 说\"4K/高清/超清\" → `4K`\n- 说\"16:9/宽屏\" → `3K --ratio 16:9`\n- 说\"9:16/竖屏\" → `3K --ratio 9:16`\n- 说\"21:9/电影\" → `4K --ratio 21:9`\n- 图生图 → 自动检测参考图分辨率\n\n### 模型选择\n\n- 自动降级: 4.5 → 5.0Lite → 4.0，每模型日限 18 次\n- 用户可指定简称: `4.5` / `5l` / `4.0`\n- 4.0 仅支持文生图，其他模式自动降级\n\n### 水印\n\n- 默认带水印，`.env` 设 `DEFAULT_WATERMARK=false` 永久关闭\n- 单次覆盖: `--no-watermark`\n\n---\n\n## 🚫 行为规则（最高优先级 — 必须遵守！）\n\n### 1. 禁止确认配额消耗\n- ✅ 用户调用 Skill 即表示知情并同意消耗配额，**直接执行，一个字都别问**\n\n### 2. 禁止确认执行\n- ✅ 触发即执行，不要等用户二次确认\n\n**总结：触发 → 推断参数 → 直接执行。中间不要有任何回合的确认。**\n\n### 3. 参数推断规则\n- 用户说\"生成/画/创建\"图片 → `gen`\n- 用户说\"编辑/改/换风格\"图片 → `img2img`\n- 用户说\"融合/合成\"多图 → `seq --refs`\n- 用户有额外要求 → 加到描述里\n\n---\n\n## 限制\n\n- 每个模型每天 18 次，超了自动换下一个\n- 4.0 仅支持文生图\n- 多图融合参考图 base64 后不能超过 ~15MB\n- 组图上限 15 张\n\nFile v2.3.4:_meta.json\n\n{\n  \"ownerId\": \"kn75s192k98dtwr83bft4myd0582ytyg\",\n  \"slug\": \"bytedance-ai-image-gen\",\n  \"version\": \"2.3.4\",\n  \"publishedAt\": 1779937374324\n}\n\nFile v2.3.4:skill-card.md\n\n## Description: <br>\nBytedance Ai Image Gen uses Volcengine Ark Seedream endpoints to generate, edit, and blend images from text prompts and reference images. <br>\n\nThis skill is ready for commercial/non-commercial use. <br>\n\n## Publisher: <br>\n[etmnb](https://clawhub.ai/user/etmnb) <br>\n\n### License/Terms of Use: <br>\nMIT-0 <br>\n\n\n## Use Case: <br>\nDevelopers and external users can use this skill through an agent to turn prompts or reference images into generated or edited image files with automatic model selection and usage limits. <br>\n\n### Deployment Geography for Use: <br>\nGlobal <br>\n\n## Known Risks and Mitigations: <br>\nRisk: The skill sends prompts and reference images to Volcengine automatically when trigger patterns match. <br>\nMitigation: Use it only with content approved for Volcengine processing, keep API credentials scoped and rotated, and avoid private photos, secrets, customer data, or proprietary prompts unless policy allows that transfer. <br>\nRisk: Generated files and plaintext prompt/reference history may remain on the local machine. <br>\nMitigation: Review or disable the local history and temporary storage behavior before using sensitive inputs, and periodically remove generated images and history files. <br>\nRisk: Broad trigger rules can consume model quota or credentials without an extra confirmation step. <br>\nMitigation: Install only in environments where automatic execution is acceptable, and review trigger patterns, quota limits, and credential access before deployment. <br>\n\n\n## Reference(s): <br>\n- [Volcengine Ark Seedream API documentation](https://www.volcengine.com/docs/82379/1541523) <br>\n- [ClawHub skill page](https://clawhub.ai/etmnb/bytedance-ai-image-gen) <br>\n\n\n## Skill Output: <br>\n**Output Type(s):** [text, markdown, shell commands, configuration, files, guidance] <br>\n**Output Format:** [Markdown guidance with shell commands, status text, and generated image file paths] <br>\n**Output Parameters:** [1D] <br>\n**Other Properties Related to Output:** [Generated images are saved locally; the skill also keeps local prompt, reference, and usage history.] <br>\n\n## Skill Version(s): <br>\n2.3.4 (source: server release metadata) <br>\n\n## Ethical Considerations: <br>\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. <br>\n\nArchive v1.0.0: 5 files, 27821 bytes\n\nFiles: _meta.json (141b), doubao_image_gen.py (71543b), image_gen_history.json (6189b), skill-card.md (2601b), SKILL.md (10874b)\n\nFile v1.0.0:SKILL.md\n\n---\nname: bytedance-ai-image-gen\ndescription: >\n  ByteDance AI Image Gen — 调用豆包/ByteDance Seedream 模型生成/编辑图片。\n  用户通过火山「协作奖励计划」免费使用顶级模型。\n  支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。\n  智能调度 4.5→5.0Lite→4.0，每模型每日18次，超限拒绝。\n  ⏱️ 生成需30-60秒，触发后请耐心等待。\nsummary: \"ByteDance AI Image Gen — 调用豆包 Seedream 生成/编辑图片，支持文生图、图生图、组图、多图融合，免费使用\"\ntags:\n  image-generation: \"2.1.0\"\n  byteimage: \"2.1.0\"\n  seedream: \"2.1.0\"\n  text-to-image: \"2.1.0\"\n  image-to-image: \"2.1.0\"\n  volcengine: \"2.1.0\"\nversion: 2.3.0\ntrigger_patterns:\n  - \"生成图片\"\n  - \"画一张图\"\n  - \"创建图片\"\n  - \"制作图片\"\n  - \"画图\"\n  - \"帮我画\"\n  - \"帮我生成\"\n  - \"我想生成\"\n  - \"搞个图\"\n  - \"来张图\"\n  - \"做张图\"\n  - \"generate image\"\n  - \"create image\"\n  - \"txt2img\"\n  - \"please generate\"\n  - \"please draw\"\n  - \"图生图\"\n  - \"图片编辑\"\n  - \"编辑图片\"\n  - \"改图\"\n  - \"换背景\"\n  - \"换风格\"\n  - \"img2img\"\n  - \"image to image\"\n  - \"edit this image\"\n  - \"edit image\"\n  - \"融合\"\n  - \"合成\"\n  - \"两张图片\"\n  - \"多张图片\"\n  - \"拼合\"\n  - \"合图\"\n  - \"blend\"\n  - \"fuse\"\n  - \"merge\"\n  - \"combine\"\n  - \"文生图\"\n  - \"生图\"\n  - \"图片生成\"\n  - \"图像生成\"\n  - \"生成图像\"\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - python\n      env:\n        - ARK_API_KEY\n        - DOUBAO_MODEL_45_ID\n        - DOUBAO_MODEL_50L_ID\n        - DOUBAO_MODEL_40_ID\n    primaryEnv: ARK_API_KEY\n    envVars:\n      - name: ARK_API_KEY\n        required: true\n        description: 火山方舟 API Key，在 console.volcengine.com/ark → API Key 管理 创建\n      - name: DOUBAO_MODEL_45_ID\n        required: true\n        description: Seedream-4.5 接入点 ID，在 console.volcengine.com/ark → 在线推理 获取\n      - name: DOUBAO_MODEL_50L_ID\n        required: true\n        description: Seedream-5.0-Lite 接入点 ID\n      - name: DOUBAO_MODEL_40_ID\n        required: true\n        description: Seedream-4.0 接入点 ID\n      - name: VOLCENGINE_ACCESS_KEY\n        required: false\n        description: 火山 IAM Access Key，在 console.volcengine.com/iam 创建，用于自动同步用量\n      - name: VOLCENGINE_SECRET_KEY\n        required: false\n        description: 火山 IAM Secret Key，与 ACCESS_KEY 配套\n    emoji: \"🎨\"\n    homepage: https://www.volcengine.com/docs/82379/1541523\n    install:\n      - kind: uv\n        package: requests\n---\n\n# ByteDance AI Image Gen — 豆包/ByteDance Seedream 图片生成\n\n调用火山方舟 Seedream [API 文档](https://www.volcengine.com/docs/82379/1541523?lang=zh) 生成/编辑图片，自动选择模型并限制用量。\n\n> ⏱️ 图片生成通常需要 **30-60 秒**，其间无额外反馈。触发后请耐心等待，不要重复发送。\n\n## ⚡ 调用方式 — 触发后必须立即执行命令\n\n> **这是操作手册，不是参考文档。检测到触发词后，直接复制对应命令执行，禁止只回复文字。**\n\n确认 Skill 目录后，先 `cd` 进去，再执行对应命令。Skill 安装在: `.openclaw/workspace/skills/bytedance-ai-image-gen/`\n\n执行成功后，脚本会输出 `✅ 生成成功!` 及本地文件路径。**你必须把以下信息告知用户**:\n- 模型、分辨率、比例\n- 本地文件路径（`文件: .../Image/xxx.png`）← 这行一定要报\n- 可以不报临时 CDN URL（会过期）\n\n### 文生图\n```bash\npython doubao_image_gen.py gen \"用户描述\" [尺寸] [--ratio 比例] [--no-watermark]\n```\n示例: `python doubao_image_gen.py gen \"洪崖洞夜景\" 4K --ratio 16:9`\n\n### 图生图\n```bash\npython doubao_image_gen.py img2img \"用户描述\" \"图片路径或URL\" [尺寸] [--no-watermark]\n```\n示例: `python doubao_image_gen.py img2img \"换成夜景风格\" \"photo.jpg\" 4K`\n\n### 组图（一次多张）\n```bash\npython doubao_image_gen.py seq \"用户描述\" [--num 数量] [尺寸] [--no-watermark]\n```\n\n### 多图融合\n```bash\npython doubao_image_gen.py seq \"融合描述\" --refs img1.jpg img2.jpg [--no-watermark]\n```\n\n⚠️ **严禁**: 只回复\"正在生成...\"而不执行命令。必须跑 `python doubao_image_gen.py ...`。\n\n---\n\n## 功能\n\n| 模式 | 说明 | 支持 |\n|------|------|------|\n| 文生单图 | 纯文本描述生成 1 张图片 | ✅ |\n| 文生组图 | 纯文本描述一次生成多张 | ✅ |\n| 图生单图 | 基于参考图编辑/风格转换 | ✅ |\n| 图生组图 | 基于参考图一次生成多张 | ✅ |\n| 多图融合 | 多张参考图融合为新图 | ✅ |\n| 用量查询 | 控制台用量同步（`python doubao_image_gen.py sync`） | ✅ |\n\n---\n\n## 🚫 行为规则（最高优先级 — 必须遵守！）\n\n**以下行为严格禁止，违反任一即为 Bug：**\n\n### 0. 默认 CLI 模式\n- ✅ 默认使用 CLI 命令：`python doubao_image_gen.py <命令> <参数...>`\n- 也可用 Python SDK 调用，但不推荐\n\n### 1. 禁止确认配额消耗\n- ✅ 用户调用 Skill 即表示知情并同意消耗配额，**直接执行，一个字都别问**\n\n### 尺寸自适应\n- ✅ 自动推断并直接执行：\n  - 没提画质 → `size=\"3K\"`（不传 ratio，API 自适应）\n  - 说\"4K/高清/超清\" → `size=\"4K\"`（不传 ratio）\n  - 说\"4K+16:9\" → `size=\"4K\", ratio=\"16:9\"`\n  - 说\"16:9/宽屏\" → `size=\"3K\", ratio=\"16:9\"`\n  - 说\"9:16/竖屏\" → `size=\"3K\", ratio=\"9:16\"`\n  - 说\"21:9/电影\" → `size=\"4K\", ratio=\"21:9\"`\n  - 明确写了尺寸 → 按指定\n- 图生图 → 自动检测参考图分辨率\n\n### 2.5 模型选择\n- 默认按优先级自动降级: 4.5 → 5.0L → 4.0\n- 用户可指定模型。简称映射：\n  - `4.5` / `4-5` / `seedream4.5` → `Doubao-Seedream-4.5`\n  - `5l` / `5.0l` / `5lite` / `seedream5lite` → `Doubao-Seedream-5.0-Lite`\n  - `4.0` / `40` / `seedream4` → `Doubao-Seedream-4.0`\n- AI 根据映射补全完整名称，传给脚本 `force_model` 参数\n- 每模型日限 18 次，超限自动跳过\n- ✅ 不询问用哪个模型，除非用户明确指定\n\n### 3. 禁止确认执行\n- ✅ 触发即执行，不要等用户二次确认\n\n**总结：触发 → 推断参数 → 直接执行。中间不要有任何回合的确认。**\n\n### 4. 水印控制\n- 默认带水印，`.env` 设 `DEFAULT_WATERMARK=false` 永久关闭\n- 单次覆盖：`--no-watermark` 强制关，不传则用默认\n\n---\n\n## 触发词\n\n### 文生图触发\n\n| 中文 | 英文 |\n|------|------|\n| `生成图片` `画一张图` `创建图片` `制作图片` | `generate image` `create image` `txt2img` |\n| `画图` `帮我画` `帮我生成` `我想生成` | `please generate` `please draw` `I want to generate` |\n| `搞个图` `来张图` `做张图` `出个图` | `draw` `make an image` `create a picture` |\n\n### 图生图 / 图片编辑触发\n\n| 中文 | 英文 |\n|------|------|\n| `图生图` `图片编辑` `编辑图片` `改图` | `img2img` `image to image` `edit this image` |\n| `换背景` `换风格` `风格转换` `图片处理` | `edit image` `modify image` `transform image` |\n\n### 多图融合 / 合成触发\n\n| 中文 | 英文 |\n|------|------|\n| `融合` `合成` `两张图片` `多张图片` | `blend` `fuse` `merge` `combine` |\n| `拼合` `合图` `融图` `图片混合` | `composite` `mix images` |\n\n### 带风格前缀\n\n`极致生成:` `快速草稿:` `艺术风格:` `4K高清:`\n\n## 工作流程\n\n调用 `doubao_image_gen.py` 脚本：\n\n1. **检测触发词** — 从用户输入提取 prompt、size、style、图生图标记\n2. **模型选择** — 按 4.5 → 5.0Lite → 4.0 降级，满 18 次跳过\n3. **尺寸自动推断** — 不再询问用户，直接按规则确定\n4. **图片规范化** — 本地路径自动 base64 编码，URL 原样通过\n5. **调用 API** — 请求火山方舟 Seedream 图片生成接口\n6. **保存图片** — 下载到 `Image/`，命名：`内容分辨率-模型-时间.png`\n7. **记录历史** — 写入 `image_gen_history.json`，7 天自动清理\n\n## 命令参考\n\n运行 `python doubao_image_gen.py --help` 查看完整参数列表和调用示例。\n\n## 模型\n\n| 模型 | 优先级 | 日限 | 文生图 | 图生图 | 组图/融合 |\n|------|--------|------|--------|--------|-----------|\n| Seedream-4.5 | 1 | 18 | ✅ | ✅ | ✅ |\n| Seedream-5.0-Lite | 2 | 18 | ✅ | ✅ | ✅ |\n| Seedream-4.0 | 3 | 18 | ✅ | ❌ 自动降级 | ❌ 自动降级 |\n\n> **已知限制**: 4.0 仅支持文生单图，其他模式自动降级到 4.5/5.0L。多图融合时注意参考图累计大小，base64 后超过 ~15MB 会被 API 拒绝。组图上限为 15 张 (参考图 + 输出)。\n\n超限返回 `{\"success\":false,\"error\":\"...\"}`。\n\n## 🎁 火山「协作奖励计划」\n\n注册火山方舟即可**免费使用** Seedream-4.5、5.0-Lite、4.0 等顶级生图模型，**无需预充值**。详情见 [火山方舟文档](https://www.volcengine.com/docs/82379/1391869)。\n\n## 配置（.env 文件）\n\n在 Skill 目录下创建 `.env` 文件：\n\n```bash\n# 必填\nARK_API_KEY=ark-your-key-here\nDOUBAO_MODEL_45_ID=ep-xxxxxxxxxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxxxxxxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxxxxxxxxx\n\n# 可选\nVOLCENGINE_ACCESS_KEY=AKLTxxx          # 控制台用量同步\nVOLCENGINE_SECRET_KEY=your-secret       # 控制台用量同步\n```\n\n> MODEL_ID 既是 API 调用 ID 也是控制台端点 ID，一个变量两用。\n\n脚本启动时自动加载 `.env`，无需手动 export。\n\n### 获取密钥\n\n- **ARK_API_KEY**: https://console.volcengine.com/ark → API Key 管理\n- **MODEL_ID**: https://console.volcengine.com/ark → 在线推理 → 点击接入点名称获取端点ID\n- **IAM 密钥**: https://console.volcengine.com/iam → 访问控制 → 密钥管理\n\n## 文件\n\n| 文件 | 说明 |\n|------|------|\n| `doubao_image_gen.py` | 核心脚本 |\n| `.env` | 环境变量配置 |\n| `image_gen_history.json` | 调用历史+计数 (7天清理) |\n| `Image/` | 生成图片保存目录 |\n| `Temp/` | 图生图本地缓存（自动创建/清理） |\n\n---\n\n## 常见错误\n\n| 错误码 | 原因 | 解决 |\n|--------|------|------|\n| `HTTP 429` | 每分钟调用频率超限 | 等待 1-2 分钟后重试 |\n| `HTTP 400` | 参数错误（尺寸/模型不支持） | 检查参数格式 |\n| `InvalidEndpointOrModel.NotFound` | 模型 ID 无效 | 检查 .env 中的 MODEL_ID |\n| `所有模型调用次数已达上限` | 三个模型都满 18 次 | 等明天重置，或查控制台确认 |\n| `图片文件不存在` | 图生图参考图路径有误 | 检查文件路径，确保文件存在 |\n| `IAM 密钥未配置` | 缺少 VOLCENGINE_ACCESS_KEY | .env 中配置 IAM 密钥 |\n\n## 💬 支持与反馈\n\n如果遇到错误或问题，请在 [clawhub.ai 技能页面](https://clawhub.ai/etmnb/bytedance-ai-image-gen) 下方留言反馈。\n\nFile v1.0.0:_meta.json\n\n{\n  \"ownerId\": \"kn75s192k98dtwr83bft4myd0582ytyg\",\n  \"slug\": \"bytedance-ai-image-gen\",\n  \"version\": \"1.0.0\",\n  \"publishedAt\": 1779935536181\n}\n\nFile v1.0.0:skill-card.md\n\n## Description: <br>\nGenerates and edits images with Doubao/ByteDance Seedream models through Volcengine, supporting text-to-image, image-to-image, batch generation, multi-image blending, and automatic model fallback. <br>\n\nThis skill is ready for commercial/non-commercial use. <br>\n\n## Publisher: <br>\n[etmnb](https://clawhub.ai/user/etmnb) <br>\n\n### License/Terms of Use: <br>\nMIT-0 <br>\n\n\n## Use Case: <br>\nExternal users and developers use this skill to generate, edit, blend, and batch-create images from prompts or reference images with ByteDance Seedream models. It is intended for agent workflows that can run a local Python CLI and return generated image file paths. <br>\n\n### Deployment Geography for Use: <br>\nGlobal <br>\n\n## Known Risks and Mitigations: <br>\nRisk: Prompts and supplied images may be sent to Volcengine/ByteDance services. <br>\nMitigation: Use only with content that is approved for external processing, and avoid private, regulated, or confidential images unless the deployment owner has approved that data flow. <br>\nRisk: Prompt and image metadata may be retained in local history files. <br>\nMitigation: Review retention settings and clean local history or temporary files when processing sensitive content. <br>\nRisk: Broad triggers can cause the skill to run without a confirmation step. <br>\nMitigation: Enable the skill only in trusted agent contexts and review trigger patterns before automatic execution is allowed. <br>\n\n\n## Reference(s): <br>\n- [ClawHub Skill Page](https://clawhub.ai/etmnb/bytedance-ai-image-gen) <br>\n- [Publisher Profile](https://clawhub.ai/user/etmnb) <br>\n- [Volcengine Seedream API Documentation](https://www.volcengine.com/docs/82379/1541523) <br>\n- [Volcengine Collaboration Rewards Documentation](https://www.volcengine.com/docs/82379/1391869) <br>\n\n\n## Skill Output: <br>\n**Output Type(s):** [Shell commands, Files, Configuration instructions, Guidance] <br>\n**Output Format:** [Markdown guidance with bash commands, JSON status responses, and local image file paths] <br>\n**Output Parameters:** [1D] <br>\n**Other Properties Related to Output:** [Generated images are saved locally; prompt and image metadata may be retained in local history.] <br>\n\n## Skill Version(s): <br>\n1.0.0 (source: server release metadata; artifact frontmatter reports 2.3.0) <br>\n\n## Ethical Considerations: <br>\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. <br>\n\nFile v1.0.0:image_gen_history.json\n\n{\r\n  \"history\": [\r\n    {\r\n      \"timestamp\": 1779774198.4971402,\r\n      \"datetime\": \"2026-05-26 13:43:18\",\r\n      \"mode\": \"text-to-image\",\r\n      \"model\": \"Doubao-Seedream-5.0-Lite\",\r\n      \"model_id\": \"ep-m-20260520180332-bw7wz\",\r\n      \"prompt\": \"画一只小舞本体\",\r\n      \"full_prompt\": \"画一只小舞本体\",\r\n      \"style\": null,\r\n      \"reference_image\": null,\r\n      \"size_requested\": \"3K\",\r\n      \"size_type\": \"preset_with_ratio\",\r\n      \"resolution\": \"4096x2304\",\r\n      \"ratio\": \"16:9\",\r\n      \"pixels\": 9437184,\r\n      \"image_url\": \"https://ark-acg-cn-beijing.tos-cn-beijing.volces.com/doubao-seedream-5-0/0217797741673267dc2a4feb4dcf772dd9b0c45c4454bcc634f44_0.jpeg?X-Tos-Algorithm=TOS4-HMAC-SHA256&X-Tos-Credential=AKLTYWJkZTExNjA1ZDUyNDc3YzhjNTM5OGIyNjBhNDcyOTQ%2F20260526%2Fcn-beijing%2Ftos%2Frequest&X-Tos-Date=20260526T054320Z&X-Tos-Expires=86400&X-Tos-Signature=7a8c151f27d7373dd1022f614866f51817b262d16f4d335e9d2b56b5d8389b1e&X-Tos-SignedHeaders=host\",\r\n      \"image_url_short\": \"https://ark-acg-cn-beijing.tos-cn-beijing.volces.com/doubao-seedream-5-0/0217797741673267dc2a4feb4dcf772dd9b0c45c4454bcc634f44_0.jpeg?X-Tos-Algorithm=TOS4-HMAC-SHA256&X-Tos-Credential=AKLTYWJkZTExN...\"\r\n    },\r\n    {\r\n      \"timestamp\": 1779785563.3068445,\r\n      \"datetime\": \"2026-05-26 16:52:43\",\r\n      \"mode\": \"text-to-image\",\r\n      \"model\": \"Doubao-Seedream-5.0-Lite\",\r\n      \"model_id\": \"ep-m-20260520180332-bw7wz\",\r\n      \"prompt\": \"深圳北站\",\r\n      \"full_prompt\": \"深圳北站\",\r\n      \"style\": null,\r\n      \"reference_image\": null,\r\n      \"size_requested\": \"3K\",\r\n      \"size_type\": \"preset_with_ratio\",\r\n      \"resolution\": \"4096x2304\",\r\n      \"ratio\": \"16:9\",\r\n      \"pixels\": 9437184,\r\n      \"image_url\": \"https://ark-acg-cn-beijing.tos-cn-beijing.volces.com/doubao-seedream-5-0/0217797855396818a2cc6176b4dc89676d1081d755bfbd1e049bd_0.jpeg?X-Tos-Algorithm=TOS4-HMAC-SHA256&X-Tos-Credential=AKLTYWJkZTExNjA1ZDUyNDc3YzhjNTM5OGIyNjBhNDcyOTQ%2F20260526%2Fcn-beijing%2Ftos%2Frequest&X-Tos-Date=20260526T085243Z&X-Tos-Expires=86400&X-Tos-Signature=e48abac38f02e7b0aa3b8f9373f6b0a5f8b32270c1b96f27e42618c47390d6ec&X-Tos-SignedHeaders=host\",\r\n      \"image_url_short\": \"https://ark-acg-cn-beijing.tos-cn-beijing.volces.com/doubao-seedream-5-0/0217797855396818a2cc6176b4dc89676d1081d755bfbd1e049bd_0.jpeg?X-Tos-Algorithm=TOS4-HMAC-SHA256&X-Tos-Credential=AKLTYWJkZTExN...\"\r\n    },\r\n    {\r\n      \"timestamp\": 1779786266.5850406,\r\n      \"datetime\": \"2026-05-26 17:04:26\",\r\n      \"mode\": \"text-to-image\",\r\n      \"model\": \"Doubao-Seedream-5.0-Lite\",\r\n      \"model_id\": \"ep-m-20260520180332-bw7wz\",\r\n      \"prompt\": \"洪崖洞\",\r\n      \"full_prompt\": \"洪崖洞\",\r\n      \"style\": null,\r\n      \"reference_image\": null,\r\n      \"size_requested\": \"3K\",\r\n      \"size_type\": \"preset_with_ratio\",\r\n      \"resolution\": \"4096x2304\",\r\n      \"ratio\": \"16:9\",\r\n      \"pixels\": 9437184,\r\n      \"image_url\": \"https://ark-acg-cn-beijing.tos-cn-beijing.volces.com/doubao-seedream-5-0/021779786243849fcaf69e354b9959077c8dee8e8ff31e04910fa_0.jpeg?X-Tos-Algorithm=TOS4-HMAC-SHA256&X-Tos-Credential=AKLTYWJkZTExNjA1ZDUyNDc3YzhjNTM5OGIyNjBhNDcyOTQ%2F20260526%2Fcn-beijing%2Ftos%2Frequest&X-Tos-Date=20260526T090427Z&X-Tos-Expires=86400&X-Tos-Signature=2c9a0439738aeb61e3a076bb7c81d028d62a048b6fedf6a3f790692036a3bcae&X-Tos-SignedHeaders=host\",\r\n      \"image_url_short\": \"https://ark-acg-cn-beijing.tos-cn-beijing.volces.com/doubao-seedream-5-0/021779786243849fcaf69e354b9959077c8dee8e8ff31e04910fa_0.jpeg?X-Tos-Algorithm=TOS4-HMAC-SHA256&X-Tos-Credential=AKLTYWJkZTExN...\"\r\n    },\r\n    {\r\n      \"timestamp\": 1779792213.658885,\r\n      \"datetime\": \"2026-05-26 18:43:33\",\r\n      \"mode\": \"text-to-image\",\r\n      \"model\": \"Doubao-Seedream-5.0-Lite\",\r\n      \"model_id\": \"ep-m-20260520180332-bw7wz\",\r\n      \"prompt\": \"拉市海\",\r\n      \"full_prompt\": \"拉市海\",\r\n      \"style\": null,\r\n      \"reference_image\": null,\r\n      \"size_requested\": \"4K\",\r\n      \"size_type\": \"preset_with_ratio\",\r\n      \"resolution\": \"4729x3546\",\r\n      \"ratio\": \"4729:3546\",\r\n      \"pixels\": 16769034,\r\n      \"image_url\": \"https://ark-acg-cn-beijing.tos-cn-beijing.volces.com/doubao-seedream-5-0/02177979217766476a6743aa0fde56972f5060a904fa21850f802_0.jpeg?X-Tos-Algorithm=TOS4-HMAC-SHA256&X-Tos-Credential=AKLTYWJkZTExNjA1ZDUyNDc3YzhjNTM5OGIyNjBhNDcyOTQ%2F20260526%2Fcn-beijing%2Ftos%2Frequest&X-Tos-Date=20260526T104334Z&X-Tos-Expires=86400&X-Tos-Signature=71ff5ee581b04dd9cc8d5e1ec424de3f1c910b4d7af1d8c7543c3e03dbc981ad&X-Tos-SignedHeaders=host\",\r\n      \"image_url_short\": \"https://ark-acg-cn-beijing.tos-cn-beijing.volces.com/doubao-seedream-5-0/02177979217766476a6743aa0fde56972f5060a904fa21850f802_0.jpeg?X-Tos-Algorithm=TOS4-HMAC-SHA256&X-Tos-Credential=AKLTYWJkZTExN...\"\r\n    },\r\n    {\r\n      \"timestamp\": 1779888361.5732768,\r\n      \"datetime\": \"2026-05-27 21:26:01\",\r\n      \"mode\": \"text-to-image\",\r\n      \"model\": \"Doubao-Seedream-4.5\",\r\n      \"model_id\": \"ep-m-20260520180218-ck8p5\",\r\n      \"prompt\": \"丽江古城夜景，16:9比例，4K\",\r\n      \"full_prompt\": \"丽江古城夜景，16:9比例，4K\",\r\n      \"style\": null,\r\n      \"reference_image\": null,\r\n      \"size_requested\": \"4K\",\r\n      \"size_type\": \"preset_with_ratio\",\r\n      \"resolution\": \"5461x3072\",\r\n      \"ratio\": \"5461:3072\",\r\n      \"pixels\": 16776192,\r\n      \"image_url\": \"https://ark-content-generation-v2-cn-beijing.tos-cn-beijing.volces.com/doubao-seedream-4-5/021779888334148e1e51997f329de504be0f95a5bc20dc73e6637_0.jpeg?X-Tos-Algorithm=TOS4-HMAC-SHA256&X-Tos-Credential=AKLTYWJkZTExNjA1ZDUyNDc3YzhjNTM5OGIyNjBhNDcyOTQ%2F20260527%2Fcn-beijing%2Ftos%2Frequest&X-Tos-Date=20260527T132601Z&X-Tos-Expires=86400&X-Tos-Signature=21487212969d03dddf4d9ac6fba230a7c298624db8fa2527f97227112f75c71c&X-Tos-SignedHeaders=host\",\r\n      \"image_url_short\": \"https://ark-content-generation-v2-cn-beijing.tos-cn-beijing.volces.com/doubao-seedream-4-5/021779888334148e1e51997f329de504be0f95a5bc20dc73e6637_0.jpeg?X-Tos-Algorithm=TOS4-HMAC-SHA256&X-Tos-Creden...\"\r\n    }\r\n  ],\r\n  \"daily_counts\": {\r\n    \"2026-05-26\": {\r\n      \"Doubao-Seedream-4.5\": 19,\r\n      \"Doubao-Seedream-5.0-Lite\": 5\r\n    },\r\n    \"2026-05-27\": {\r\n      \"Doubao-Seedream-4.5\": 1\r\n    }\r\n  }\r\n}\n\nArchive v2.3.3: 3 files, 23371 bytes\n\nFiles: _meta.json (141b), doubao_image_gen.py (71543b), SKILL.md (6839b)\n\nFile v2.3.3:SKILL.md\n\n---\nname: bytedance-ai-image-gen\ndescription: >\n  ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。\n  支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。\n  参与火山协作奖励计划免费使用顶级生图模型。\n  智能调度模型，每模型每日18次，超限拒绝。\n  ⏱️ 生成需30-60秒，触发后请耐心等待。\nsummary: \"ByteDance AI Image Gen — 调用豆包 Seedream 生成/编辑图片，支持文生图、图生图、组图、多图融合，参与火山协作奖励计划免费使用\"\ntags:\n  image-generation: \"2.1.0\"\n  byteimage: \"2.1.0\"\n  seedream: \"2.1.0\"\n  text-to-image: \"2.1.0\"\n  image-to-image: \"2.1.0\"\n  volcengine: \"2.1.0\"\nversion: 2.3.0\ntrigger_patterns:\n  - \"生成图片\"\n  - \"画一张图\"\n  - \"创建图片\"\n  - \"制作图片\"\n  - \"画图\"\n  - \"帮我画\"\n  - \"帮我生成\"\n  - \"我想生成\"\n  - \"搞个图\"\n  - \"来张图\"\n  - \"做张图\"\n  - \"generate image\"\n  - \"create image\"\n  - \"txt2img\"\n  - \"please generate\"\n  - \"please draw\"\n  - \"图生图\"\n  - \"图片编辑\"\n  - \"编辑图片\"\n  - \"改图\"\n  - \"换背景\"\n  - \"换风格\"\n  - \"img2img\"\n  - \"image to image\"\n  - \"edit this image\"\n  - \"edit image\"\n  - \"融合\"\n  - \"合成\"\n  - \"两张图片\"\n  - \"多张图片\"\n  - \"拼合\"\n  - \"合图\"\n  - \"blend\"\n  - \"fuse\"\n  - \"merge\"\n  - \"combine\"\n  - \"文生图\"\n  - \"生图\"\n  - \"图片生成\"\n  - \"图像生成\"\n  - \"生成图像\"\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - python\n      env:\n        - ARK_API_KEY\n        - DOUBAO_MODEL_45_ID\n        - DOUBAO_MODEL_50L_ID\n        - DOUBAO_MODEL_40_ID\n    primaryEnv: ARK_API_KEY\n    envVars:\n      - name: ARK_API_KEY\n        required: true\n        description: 火山方舟 API Key，在 console.volcengine.com/ark → API Key 管理 创建\n      - name: DOUBAO_MODEL_45_ID\n        required: true\n        description: Seedream-4.5 接入点 ID，在 console.volcengine.com/ark → 在线推理 获取\n      - name: DOUBAO_MODEL_50L_ID\n        required: true\n        description: Seedream-5.0-Lite 接入点 ID\n      - name: DOUBAO_MODEL_40_ID\n        required: true\n        description: Seedream-4.0 接入点 ID\n      - name: VOLCENGINE_ACCESS_KEY\n        required: false\n        description: 火山 IAM Access Key，在 console.volcengine.com/iam 创建，用于自动同步用量\n      - name: VOLCENGINE_SECRET_KEY\n        required: false\n        description: 火山 IAM Secret Key，与 ACCESS_KEY 配套\n    emoji: \"🎨\"\n    homepage: https://www.volcengine.com/docs/82379/1541523\n    install:\n      - kind: uv\n        package: requests\n---\n\n# ByteDance AI Image Gen — 豆包 Seedream 图片生成\n\n调用火山方舟 Seedream [API 文档](https://www.volcengine.com/docs/82379/1541523?lang=zh) 生成/编辑图片，自动选择模型并限制用量。参与火山协作奖励计划免费使用顶级生图模型。\n\n> ⚠️ **配置只需一次！** 如果 `.env` 文件已存在且包含 `ARK_API_KEY` 和 3 个模型 ID，说明已配置过，**直接跳到\"调用方式\"执行命令，不要重新配置。**\n\n## 🚀 首次配置\n\n### 1. 获取 API Key\n\n1. 打开 https://console.volcengine.com/ark 注册/登录\n2. 左侧菜单 → **API Key 管理** → 创建 API Key → 复制保存\n\n### 2. 创建模型接入点\n\n在同一个控制台，左侧菜单 → **在线推理** → 创建推理接入点，选以下 3 个模型：\n\n| 环境变量名 | 模型 | 说明 |\n|-----------|------|------|\n| `DOUBAO_MODEL_45_ID` | Seedream-4.5 | 主力模型，所有模式优先 |\n| `DOUBAO_MODEL_50L_ID` | Seedream-5.0-Lite | 轻量备选 |\n| `DOUBAO_MODEL_40_ID` | Seedream-4.0 | 仅支持文生图 |\n\n每个接入点创建后会得到一个 `ep-xxxxx` 格式的 ID，复制保存。\n\n### 3. 配置 .env 文件\n\n在 Skill 目录下创建 `.env` 文件，填入你的 Key 和接入点 ID：\n\n```bash\nARK_API_KEY=你的API Key\nDOUBAO_MODEL_45_ID=ep-xxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxx\n```\n\n> 可选：加上 IAM 密钥可自动同步控制台用量\n> ```bash\n> VOLCENGINE_ACCESS_KEY=你的Access Key\n> VOLCENGINE_SECRET_KEY=你的Secret Key\n> ```\n\n### 4. 测试\n\n```bash\npython doubao_image_gen.py --help\n```\n\n不报错，就配置好了。\n\n---\n\n## ⚡ 调用方式 — 触发后必须立即执行命令\n\n> **这是操作手册，不是参考文档。检测到触发词后，直接复制对应命令执行，禁止只回复文字。**\n\n确认 Skill 目录后，先 `cd` 进去，再执行对应命令。Skill 安装在: `.openclaw/workspace/skills/bytedance-ai-image-gen/`\n\n执行成功后，脚本会输出 `✅ 生成成功!` 及本地文件路径。**你必须把文件路径告知用户**。\n\n⚠️ **严禁**: 只回复\"正在生成...\"而不执行命令。必须跑 `python doubao_image_gen.py ...`。\n\n### 命令\n\n| 命令 | 用途 | 示例 |\n|------|------|------|\n| `gen \"描述\" [尺寸] [--ratio 比例]` | 文生图 | `gen \"洪崖洞夜景\" 4K --ratio 16:9` |\n| `img2img \"描述\" \"图片\" [尺寸]` | 图生图 | `img2img \"换成夜景\" \"photo.jpg\" 4K` |\n| `seq \"描述\" [--num 数量]` | 组图 | `seq \"猫咪\" --num 3` |\n| `seq \"描述\" --refs img1 img2` | 多图融合 | `seq \"融合\" --refs a.jpg b.jpg` |\n| `sync` | 同步控制台用量 | |\n\n### 参数\n\n| 参数 | 说明 |\n|------|------|\n| `尺寸` | `3K`(默认) / `4K`(高清) / 自定义 |\n| `--ratio` | `16:9`(宽屏) / `9:16`(竖屏) / `21:9`(电影) |\n| `--no-watermark` | 去水印 |\n| `--num` | 组图数量 |\n\n### 尺寸推断规则\n\n- 没提画质 → `3K`\n- 说\"4K/高清/超清\" → `4K`\n- 说\"16:9/宽屏\" → `3K --ratio 16:9`\n- 说\"9:16/竖屏\" → `3K --ratio 9:16`\n- 说\"21:9/电影\" → `4K --ratio 21:9`\n- 图生图 → 自动检测参考图分辨率\n\n### 模型选择\n\n- 自动降级: 4.5 → 5.0Lite → 4.0，每模型日限 18 次\n- 用户可指定简称: `4.5` / `5l` / `4.0`\n- 4.0 仅支持文生图，其他模式自动降级\n\n### 水印\n\n- 默认带水印，`.env` 设 `DEFAULT_WATERMARK=false` 永久关闭\n- 单次覆盖: `--no-watermark`\n\n---\n\n## 🚫 行为规则（最高优先级 — 必须遵守！）\n\n### 1. 禁止确认配额消耗\n- ✅ 用户调用 Skill 即表示知情并同意消耗配额，**直接执行，一个字都别问**\n\n### 2. 禁止确认执行\n- ✅ 触发即执行，不要等用户二次确认\n\n**总结：触发 → 推断参数 → 直接执行。中间不要有任何回合的确认。**\n\n### 3. 参数推断规则\n- 用户说\"生成/画/创建\"图片 → `gen`\n- 用户说\"编辑/改/换风格\"图片 → `img2img`\n- 用户说\"融合/合成\"多图 → `seq --refs`\n- 用户有额外要求 → 加到描述里\n\n---\n\n## 限制\n\n- 每个模型每天 18 次，超了自动换下一个\n- 4.0 仅支持文生图\n- 多图融合参考图 base64 后不能超过 ~15MB\n- 组图上限 15 张\n\nFile v2.3.3:_meta.json\n\n{\n  \"ownerId\": \"kn75s192k98dtwr83bft4myd0582ytyg\",\n  \"slug\": \"bytedance-ai-image-gen\",\n  \"version\": \"2.3.3\",\n  \"publishedAt\": 1779907126329\n}\n\nArchive v2.3.2: 3 files, 23371 bytes\n\nFiles: _meta.json (141b), doubao_image_gen.py (71543b), SKILL.md (6839b)\n\nFile v2.3.2:SKILL.md\n\n---\nname: bytedance-ai-image-gen\ndescription: >\n  ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。\n  支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。\n  参与火山协作奖励计划免费使用顶级生图模型。\n  智能调度模型，每模型每日18次，超限拒绝。\n  ⏱️ 生成需30-60秒，触发后请耐心等待。\nsummary: \"ByteDance AI Image Gen — 调用豆包 Seedream 生成/编辑图片，支持文生图、图生图、组图、多图融合，参与火山协作奖励计划免费使用\"\ntags:\n  image-generation: \"2.1.0\"\n  byteimage: \"2.1.0\"\n  seedream: \"2.1.0\"\n  text-to-image: \"2.1.0\"\n  image-to-image: \"2.1.0\"\n  volcengine: \"2.1.0\"\nversion: 2.3.0\ntrigger_patterns:\n  - \"生成图片\"\n  - \"画一张图\"\n  - \"创建图片\"\n  - \"制作图片\"\n  - \"画图\"\n  - \"帮我画\"\n  - \"帮我生成\"\n  - \"我想生成\"\n  - \"搞个图\"\n  - \"来张图\"\n  - \"做张图\"\n  - \"generate image\"\n  - \"create image\"\n  - \"txt2img\"\n  - \"please generate\"\n  - \"please draw\"\n  - \"图生图\"\n  - \"图片编辑\"\n  - \"编辑图片\"\n  - \"改图\"\n  - \"换背景\"\n  - \"换风格\"\n  - \"img2img\"\n  - \"image to image\"\n  - \"edit this image\"\n  - \"edit image\"\n  - \"融合\"\n  - \"合成\"\n  - \"两张图片\"\n  - \"多张图片\"\n  - \"拼合\"\n  - \"合图\"\n  - \"blend\"\n  - \"fuse\"\n  - \"merge\"\n  - \"combine\"\n  - \"文生图\"\n  - \"生图\"\n  - \"图片生成\"\n  - \"图像生成\"\n  - \"生成图像\"\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - python\n      env:\n        - ARK_API_KEY\n        - DOUBAO_MODEL_45_ID\n        - DOUBAO_MODEL_50L_ID\n        - DOUBAO_MODEL_40_ID\n    primaryEnv: ARK_API_KEY\n    envVars:\n      - name: ARK_API_KEY\n        required: true\n        description: 火山方舟 API Key，在 console.volcengine.com/ark → API Key 管理 创建\n      - name: DOUBAO_MODEL_45_ID\n        required: true\n        description: Seedream-4.5 接入点 ID，在 console.volcengine.com/ark → 在线推理 获取\n      - name: DOUBAO_MODEL_50L_ID\n        required: true\n        description: Seedream-5.0-Lite 接入点 ID\n      - name: DOUBAO_MODEL_40_ID\n        required: true\n        description: Seedream-4.0 接入点 ID\n      - name: VOLCENGINE_ACCESS_KEY\n        required: false\n        description: 火山 IAM Access Key，在 console.volcengine.com/iam 创建，用于自动同步用量\n      - name: VOLCENGINE_SECRET_KEY\n        required: false\n        description: 火山 IAM Secret Key，与 ACCESS_KEY 配套\n    emoji: \"🎨\"\n    homepage: https://www.volcengine.com/docs/82379/1541523\n    install:\n      - kind: uv\n        package: requests\n---\n\n# ByteDance AI Image Gen — 豆包 Seedream 图片生成\n\n调用火山方舟 Seedream [API 文档](https://www.volcengine.com/docs/82379/1541523?lang=zh) 生成/编辑图片，自动选择模型并限制用量。参与火山协作奖励计划免费使用顶级生图模型。\n\n> ⚠️ **配置只需一次！** 如果 `.env` 文件已存在且包含 `ARK_API_KEY` 和 3 个模型 ID，说明已配置过，**直接跳到\"调用方式\"执行命令，不要重新配置。**\n\n## 🚀 首次配置\n\n### 1. 获取 API Key\n\n1. 打开 https://console.volcengine.com/ark 注册/登录\n2. 左侧菜单 → **API Key 管理** → 创建 API Key → 复制保存\n\n### 2. 创建模型接入点\n\n在同一个控制台，左侧菜单 → **在线推理** → 创建推理接入点，选以下 3 个模型：\n\n| 环境变量名 | 模型 | 说明 |\n|-----------|------|------|\n| `DOUBAO_MODEL_45_ID` | Seedream-4.5 | 主力模型，所有模式优先 |\n| `DOUBAO_MODEL_50L_ID` | Seedream-5.0-Lite | 轻量备选 |\n| `DOUBAO_MODEL_40_ID` | Seedream-4.0 | 仅支持文生图 |\n\n每个接入点创建后会得到一个 `ep-xxxxx` 格式的 ID，复制保存。\n\n### 3. 配置 .env 文件\n\n在 Skill 目录下创建 `.env` 文件，填入你的 Key 和接入点 ID：\n\n```bash\nARK_API_KEY=你的API Key\nDOUBAO_MODEL_45_ID=ep-xxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxx\n```\n\n> 可选：加上 IAM 密钥可自动同步控制台用量\n> ```bash\n> VOLCENGINE_ACCESS_KEY=你的Access Key\n> VOLCENGINE_SECRET_KEY=你的Secret Key\n> ```\n\n### 4. 测试\n\n```bash\npython doubao_image_gen.py --help\n```\n\n不报错，就配置好了。\n\n---\n\n## ⚡ 调用方式 — 触发后必须立即执行命令\n\n> **这是操作手册，不是参考文档。检测到触发词后，直接复制对应命令执行，禁止只回复文字。**\n\n确认 Skill 目录后，先 `cd` 进去，再执行对应命令。Skill 安装在: `.openclaw/workspace/skills/bytedance-ai-image-gen/`\n\n执行成功后，脚本会输出 `✅ 生成成功!` 及本地文件路径。**你必须把文件路径告知用户**。\n\n⚠️ **严禁**: 只回复\"正在生成...\"而不执行命令。必须跑 `python doubao_image_gen.py ...`。\n\n### 命令\n\n| 命令 | 用途 | 示例 |\n|------|------|------|\n| `gen \"描述\" [尺寸] [--ratio 比例]` | 文生图 | `gen \"洪崖洞夜景\" 4K --ratio 16:9` |\n| `img2img \"描述\" \"图片\" [尺寸]` | 图生图 | `img2img \"换成夜景\" \"photo.jpg\" 4K` |\n| `seq \"描述\" [--num 数量]` | 组图 | `seq \"猫咪\" --num 3` |\n| `seq \"描述\" --refs img1 img2` | 多图融合 | `seq \"融合\" --refs a.jpg b.jpg` |\n| `sync` | 同步控制台用量 | |\n\n### 参数\n\n| 参数 | 说明 |\n|------|------|\n| `尺寸` | `3K`(默认) / `4K`(高清) / 自定义 |\n| `--ratio` | `16:9`(宽屏) / `9:16`(竖屏) / `21:9`(电影) |\n| `--no-watermark` | 去水印 |\n| `--num` | 组图数量 |\n\n### 尺寸推断规则\n\n- 没提画质 → `3K`\n- 说\"4K/高清/超清\" → `4K`\n- 说\"16:9/宽屏\" → `3K --ratio 16:9`\n- 说\"9:16/竖屏\" → `3K --ratio 9:16`\n- 说\"21:9/电影\" → `4K --ratio 21:9`\n- 图生图 → 自动检测参考图分辨率\n\n### 模型选择\n\n- 自动降级: 4.5 → 5.0Lite → 4.0，每模型日限 18 次\n- 用户可指定简称: `4.5` / `5l` / `4.0`\n- 4.0 仅支持文生图，其他模式自动降级\n\n### 水印\n\n- 默认带水印，`.env` 设 `DEFAULT_WATERMARK=false` 永久关闭\n- 单次覆盖: `--no-watermark`\n\n---\n\n## 🚫 行为规则（最高优先级 — 必须遵守！）\n\n### 1. 禁止确认配额消耗\n- ✅ 用户调用 Skill 即表示知情并同意消耗配额，**直接执行，一个字都别问**\n\n### 2. 禁止确认执行\n- ✅ 触发即执行，不要等用户二次确认\n\n**总结：触发 → 推断参数 → 直接执行。中间不要有任何回合的确认。**\n\n### 3. 参数推断规则\n- 用户说\"生成/画/创建\"图片 → `gen`\n- 用户说\"编辑/改/换风格\"图片 → `img2img`\n- 用户说\"融合/合成\"多图 → `seq --refs`\n- 用户有额外要求 → 加到描述里\n\n---\n\n## 限制\n\n- 每个模型每天 18 次，超了自动换下一个\n- 4.0 仅支持文生图\n- 多图融合参考图 base64 后不能超过 ~15MB\n- 组图上限 15 张\n\nFile v2.3.2:_meta.json\n\n{\n  \"ownerId\": \"kn75s192k98dtwr83bft4myd0582ytyg\",\n  \"slug\": \"bytedance-ai-image-gen\",\n  \"version\": \"2.3.2\",\n  \"publishedAt\": 1779906570389\n}\n\nArchive v2.3.1: 3 files, 23371 bytes\n\nFiles: _meta.json (141b), doubao_image_gen.py (71543b), SKILL.md (6839b)\n\nFile v2.3.1:SKILL.md\n\n---\nname: bytedance-ai-image-gen\ndescription: >\n  ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。\n  支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。\n  参与火山协作奖励计划免费使用顶级生图模型。\n  智能调度模型，每模型每日18次，超限拒绝。\n  ⏱️ 生成需30-60秒，触发后请耐心等待。\nsummary: \"ByteDance AI Image Gen — 调用豆包 Seedream 生成/编辑图片，支持文生图、图生图、组图、多图融合，参与火山协作奖励计划免费使用\"\ntags:\n  image-generation: \"2.1.0\"\n  byteimage: \"2.1.0\"\n  seedream: \"2.1.0\"\n  text-to-image: \"2.1.0\"\n  image-to-image: \"2.1.0\"\n  volcengine: \"2.1.0\"\nversion: 2.3.0\ntrigger_patterns:\n  - \"生成图片\"\n  - \"画一张图\"\n  - \"创建图片\"\n  - \"制作图片\"\n  - \"画图\"\n  - \"帮我画\"\n  - \"帮我生成\"\n  - \"我想生成\"\n  - \"搞个图\"\n  - \"来张图\"\n  - \"做张图\"\n  - \"generate image\"\n  - \"create image\"\n  - \"txt2img\"\n  - \"please generate\"\n  - \"please draw\"\n  - \"图生图\"\n  - \"图片编辑\"\n  - \"编辑图片\"\n  - \"改图\"\n  - \"换背景\"\n  - \"换风格\"\n  - \"img2img\"\n  - \"image to image\"\n  - \"edit this image\"\n  - \"edit image\"\n  - \"融合\"\n  - \"合成\"\n  - \"两张图片\"\n  - \"多张图片\"\n  - \"拼合\"\n  - \"合图\"\n  - \"blend\"\n  - \"fuse\"\n  - \"merge\"\n  - \"combine\"\n  - \"文生图\"\n  - \"生图\"\n  - \"图片生成\"\n  - \"图像生成\"\n  - \"生成图像\"\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - python\n      env:\n        - ARK_API_KEY\n        - DOUBAO_MODEL_45_ID\n        - DOUBAO_MODEL_50L_ID\n        - DOUBAO_MODEL_40_ID\n    primaryEnv: ARK_API_KEY\n    envVars:\n      - name: ARK_API_KEY\n        required: true\n        description: 火山方舟 API Key，在 console.volcengine.com/ark → API Key 管理 创建\n      - name: DOUBAO_MODEL_45_ID\n        required: true\n        description: Seedream-4.5 接入点 ID，在 console.volcengine.com/ark → 在线推理 获取\n      - name: DOUBAO_MODEL_50L_ID\n        required: true\n        description: Seedream-5.0-Lite 接入点 ID\n      - name: DOUBAO_MODEL_40_ID\n        required: true\n        description: Seedream-4.0 接入点 ID\n      - name: VOLCENGINE_ACCESS_KEY\n        required: false\n        description: 火山 IAM Access Key，在 console.volcengine.com/iam 创建，用于自动同步用量\n      - name: VOLCENGINE_SECRET_KEY\n        required: false\n        description: 火山 IAM Secret Key，与 ACCESS_KEY 配套\n    emoji: \"🎨\"\n    homepage: https://www.volcengine.com/docs/82379/1541523\n    install:\n      - kind: uv\n        package: requests\n---\n\n# ByteDance AI Image Gen — 豆包 Seedream 图片生成\n\n调用火山方舟 Seedream [API 文档](https://www.volcengine.com/docs/82379/1541523?lang=zh) 生成/编辑图片，自动选择模型并限制用量。参与火山协作奖励计划免费使用顶级生图模型。\n\n> ⚠️ **配置只需一次！** 如果 `.env` 文件已存在且包含 `ARK_API_KEY` 和 3 个模型 ID，说明已配置过，**直接跳到\"调用方式\"执行命令，不要重新配置。**\n\n## 🚀 首次配置\n\n### 1. 获取 API Key\n\n1. 打开 https://console.volcengine.com/ark 注册/登录\n2. 左侧菜单 → **API Key 管理** → 创建 API Key → 复制保存\n\n### 2. 创建模型接入点\n\n在同一个控制台，左侧菜单 → **在线推理** → 创建推理接入点，选以下 3 个模型：\n\n| 环境变量名 | 模型 | 说明 |\n|-----------|------|------|\n| `DOUBAO_MODEL_45_ID` | Seedream-4.5 | 主力模型，所有模式优先 |\n| `DOUBAO_MODEL_50L_ID` | Seedream-5.0-Lite | 轻量备选 |\n| `DOUBAO_MODEL_40_ID` | Seedream-4.0 | 仅支持文生图 |\n\n每个接入点创建后会得到一个 `ep-xxxxx` 格式的 ID，复制保存。\n\n### 3. 配置 .env 文件\n\n在 Skill 目录下创建 `.env` 文件，填入你的 Key 和接入点 ID：\n\n```bash\nARK_API_KEY=你的API Key\nDOUBAO_MODEL_45_ID=ep-xxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxx\n```\n\n> 可选：加上 IAM 密钥可自动同步控制台用量\n> ```bash\n> VOLCENGINE_ACCESS_KEY=你的Access Key\n> VOLCENGINE_SECRET_KEY=你的Secret Key\n> ```\n\n### 4. 测试\n\n```bash\npython doubao_image_gen.py --help\n```\n\n不报错，就配置好了。\n\n---\n\n## ⚡ 调用方式 — 触发后必须立即执行命令\n\n> **这是操作手册，不是参考文档。检测到触发词后，直接复制对应命令执行，禁止只回复文字。**\n\n确认 Skill 目录后，先 `cd` 进去，再执行对应命令。Skill 安装在: `.openclaw/workspace/skills/bytedance-ai-image-gen/`\n\n执行成功后，脚本会输出 `✅ 生成成功!` 及本地文件路径。**你必须把文件路径告知用户**。\n\n⚠️ **严禁**: 只回复\"正在生成...\"而不执行命令。必须跑 `python doubao_image_gen.py ...`。\n\n### 命令\n\n| 命令 | 用途 | 示例 |\n|------|------|------|\n| `gen \"描述\" [尺寸] [--ratio 比例]` | 文生图 | `gen \"洪崖洞夜景\" 4K --ratio 16:9` |\n| `img2img \"描述\" \"图片\" [尺寸]` | 图生图 | `img2img \"换成夜景\" \"photo.jpg\" 4K` |\n| `seq \"描述\" [--num 数量]` | 组图 | `seq \"猫咪\" --num 3` |\n| `seq \"描述\" --refs img1 img2` | 多图融合 | `seq \"融合\" --refs a.jpg b.jpg` |\n| `sync` | 同步控制台用量 | |\n\n### 参数\n\n| 参数 | 说明 |\n|------|------|\n| `尺寸` | `3K`(默认) / `4K`(高清) / 自定义 |\n| `--ratio` | `16:9`(宽屏) / `9:16`(竖屏) / `21:9`(电影) |\n| `--no-watermark` | 去水印 |\n| `--num` | 组图数量 |\n\n### 尺寸推断规则\n\n- 没提画质 → `3K`\n- 说\"4K/高清/超清\" → `4K`\n- 说\"16:9/宽屏\" → `3K --ratio 16:9`\n- 说\"9:16/竖屏\" → `3K --ratio 9:16`\n- 说\"21:9/电影\" → `4K --ratio 21:9`\n- 图生图 → 自动检测参考图分辨率\n\n### 模型选择\n\n- 自动降级: 4.5 → 5.0Lite → 4.0，每模型日限 18 次\n- 用户可指定简称: `4.5` / `5l` / `4.0`\n- 4.0 仅支持文生图，其他模式自动降级\n\n### 水印\n\n- 默认带水印，`.env` 设 `DEFAULT_WATERMARK=false` 永久关闭\n- 单次覆盖: `--no-watermark`\n\n---\n\n## 🚫 行为规则（最高优先级 — 必须遵守！）\n\n### 1. 禁止确认配额消耗\n- ✅ 用户调用 Skill 即表示知情并同意消耗配额，**直接执行，一个字都别问**\n\n### 2. 禁止确认执行\n- ✅ 触发即执行，不要等用户二次确认\n\n**总结：触发 → 推断参数 → 直接执行。中间不要有任何回合的确认。**\n\n### 3. 参数推断规则\n- 用户说\"生成/画/创建\"图片 → `gen`\n- 用户说\"编辑/改/换风格\"图片 → `img2img`\n- 用户说\"融合/合成\"多图 → `seq --refs`\n- 用户有额外要求 → 加到描述里\n\n---\n\n## 限制\n\n- 每个模型每天 18 次，超了自动换下一个\n- 4.0 仅支持文生图\n- 多图融合参考图 base64 后不能超过 ~15MB\n- 组图上限 15 张\n\nFile v2.3.1:_meta.json\n\n{\n  \"ownerId\": \"kn75s192k98dtwr83bft4myd0582ytyg\",\n  \"slug\": \"bytedance-ai-image-gen\",\n  \"version\": \"2.3.1\",\n  \"publishedAt\": 1779906445381\n}\n\nArchive v2.4.0: 4 files, 24627 bytes\n\nFiles: _meta.json (141b), doubao_image_gen.py (71543b), skill-card.md (2389b), SKILL.md (6839b)\n\nFile v2.4.0:SKILL.md\n\n---\nname: bytedance-ai-image-gen\ndescription: >\n  ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。\n  支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。\n  参与火山协作奖励计划免费使用顶级生图模型。\n  智能调度模型，每模型每日18次，超限拒绝。\n  ⏱️ 生成需30-60秒，触发后请耐心等待。\nsummary: \"ByteDance AI Image Gen — 调用豆包 Seedream 生成/编辑图片，支持文生图、图生图、组图、多图融合，参与火山协作奖励计划免费使用\"\ntags:\n  image-generation: \"2.1.0\"\n  byteimage: \"2.1.0\"\n  seedream: \"2.1.0\"\n  text-to-image: \"2.1.0\"\n  image-to-image: \"2.1.0\"\n  volcengine: \"2.1.0\"\nversion: 2.3.0\ntrigger_patterns:\n  - \"生成图片\"\n  - \"画一张图\"\n  - \"创建图片\"\n  - \"制作图片\"\n  - \"画图\"\n  - \"帮我画\"\n  - \"帮我生成\"\n  - \"我想生成\"\n  - \"搞个图\"\n  - \"来张图\"\n  - \"做张图\"\n  - \"generate image\"\n  - \"create image\"\n  - \"txt2img\"\n  - \"please generate\"\n  - \"please draw\"\n  - \"图生图\"\n  - \"图片编辑\"\n  - \"编辑图片\"\n  - \"改图\"\n  - \"换背景\"\n  - \"换风格\"\n  - \"img2img\"\n  - \"image to image\"\n  - \"edit this image\"\n  - \"edit image\"\n  - \"融合\"\n  - \"合成\"\n  - \"两张图片\"\n  - \"多张图片\"\n  - \"拼合\"\n  - \"合图\"\n  - \"blend\"\n  - \"fuse\"\n  - \"merge\"\n  - \"combine\"\n  - \"文生图\"\n  - \"生图\"\n  - \"图片生成\"\n  - \"图像生成\"\n  - \"生成图像\"\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - python\n      env:\n        - ARK_API_KEY\n        - DOUBAO_MODEL_45_ID\n        - DOUBAO_MODEL_50L_ID\n        - DOUBAO_MODEL_40_ID\n    primaryEnv: ARK_API_KEY\n    envVars:\n      - name: ARK_API_KEY\n        required: true\n        description: 火山方舟 API Key，在 console.volcengine.com/ark → API Key 管理 创建\n      - name: DOUBAO_MODEL_45_ID\n        required: true\n        description: Seedream-4.5 接入点 ID，在 console.volcengine.com/ark → 在线推理 获取\n      - name: DOUBAO_MODEL_50L_ID\n        required: true\n        description: Seedream-5.0-Lite 接入点 ID\n      - name: DOUBAO_MODEL_40_ID\n        required: true\n        description: Seedream-4.0 接入点 ID\n      - name: VOLCENGINE_ACCESS_KEY\n        required: false\n        description: 火山 IAM Access Key，在 console.volcengine.com/iam 创建，用于自动同步用量\n      - name: VOLCENGINE_SECRET_KEY\n        required: false\n        description: 火山 IAM Secret Key，与 ACCESS_KEY 配套\n    emoji: \"🎨\"\n    homepage: https://www.volcengine.com/docs/82379/1541523\n    install:\n      - kind: uv\n        package: requests\n---\n\n# ByteDance AI Image Gen — 豆包 Seedream 图片生成\n\n调用火山方舟 Seedream [API 文档](https://www.volcengine.com/docs/82379/1541523?lang=zh) 生成/编辑图片，自动选择模型并限制用量。参与火山协作奖励计划免费使用顶级生图模型。\n\n> ⚠️ **配置只需一次！** 如果 `.env` 文件已存在且包含 `ARK_API_KEY` 和 3 个模型 ID，说明已配置过，**直接跳到\"调用方式\"执行命令，不要重新配置。**\n\n## 🚀 首次配置\n\n### 1. 获取 API Key\n\n1. 打开 https://console.volcengine.com/ark 注册/登录\n2. 左侧菜单 → **API Key 管理** → 创建 API Key → 复制保存\n\n### 2. 创建模型接入点\n\n在同一个控制台，左侧菜单 → **在线推理** → 创建推理接入点，选以下 3 个模型：\n\n| 环境变量名 | 模型 | 说明 |\n|-----------|------|------|\n| `DOUBAO_MODEL_45_ID` | Seedream-4.5 | 主力模型，所有模式优先 |\n| `DOUBAO_MODEL_50L_ID` | Seedream-5.0-Lite | 轻量备选 |\n| `DOUBAO_MODEL_40_ID` | Seedream-4.0 | 仅支持文生图 |\n\n每个接入点创建后会得到一个 `ep-xxxxx` 格式的 ID，复制保存。\n\n### 3. 配置 .env 文件\n\n在 Skill 目录下创建 `.env` 文件，填入你的 Key 和接入点 ID：\n\n```bash\nARK_API_KEY=你的API Key\nDOUBAO_MODEL_45_ID=ep-xxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxx\n```\n\n> 可选：加上 IAM 密钥可自动同步控制台用量\n> ```bash\n> VOLCENGINE_ACCESS_KEY=你的Access Key\n> VOLCENGINE_SECRET_KEY=你的Secret Key\n> ```\n\n### 4. 测试\n\n```bash\npython doubao_image_gen.py --help\n```\n\n不报错，就配置好了。\n\n---\n\n## ⚡ 调用方式 — 触发后必须立即执行命令\n\n> **这是操作手册，不是参考文档。检测到触发词后，直接复制对应命令执行，禁止只回复文字。**\n\n确认 Skill 目录后，先 `cd` 进去，再执行对应命令。Skill 安装在: `.openclaw/workspace/skills/bytedance-ai-image-gen/`\n\n执行成功后，脚本会输出 `✅ 生成成功!` 及本地文件路径。**你必须把文件路径告知用户**。\n\n⚠️ **严禁**: 只回复\"正在生成...\"而不执行命令。必须跑 `python doubao_image_gen.py ...`。\n\n### 命令\n\n| 命令 | 用途 | 示例 |\n|------|------|------|\n| `gen \"描述\" [尺寸] [--ratio 比例]` | 文生图 | `gen \"洪崖洞夜景\" 4K --ratio 16:9` |\n| `img2img \"描述\" \"图片\" [尺寸]` | 图生图 | `img2img \"换成夜景\" \"photo.jpg\" 4K` |\n| `seq \"描述\" [--num 数量]` | 组图 | `seq \"猫咪\" --num 3` |\n| `seq \"描述\" --refs img1 img2` | 多图融合 | `seq \"融合\" --refs a.jpg b.jpg` |\n| `sync` | 同步控制台用量 | |\n\n### 参数\n\n| 参数 | 说明 |\n|------|------|\n| `尺寸` | `3K`(默认) / `4K`(高清) / 自定义 |\n| `--ratio` | `16:9`(宽屏) / `9:16`(竖屏) / `21:9`(电影) |\n| `--no-watermark` | 去水印 |\n| `--num` | 组图数量 |\n\n### 尺寸推断规则\n\n- 没提画质 → `3K`\n- 说\"4K/高清/超清\" → `4K`\n- 说\"16:9/宽屏\" → `3K --ratio 16:9`\n- 说\"9:16/竖屏\" → `3K --ratio 9:16`\n- 说\"21:9/电影\" → `4K --ratio 21:9`\n- 图生图 → 自动检测参考图分辨率\n\n### 模型选择\n\n- 自动降级: 4.5 → 5.0Lite → 4.0，每模型日限 18 次\n- 用户可指定简称: `4.5` / `5l` / `4.0`\n- 4.0 仅支持文生图，其他模式自动降级\n\n### 水印\n\n- 默认带水印，`.env` 设 `DEFAULT_WATERMARK=false` 永久关闭\n- 单次覆盖: `--no-watermark`\n\n---\n\n## 🚫 行为规则（最高优先级 — 必须遵守！）\n\n### 1. 禁止确认配额消耗\n- ✅ 用户调用 Skill 即表示知情并同意消耗配额，**直接执行，一个字都别问**\n\n### 2. 禁止确认执行\n- ✅ 触发即执行，不要等用户二次确认\n\n**总结：触发 → 推断参数 → 直接执行。中间不要有任何回合的确认。**\n\n### 3. 参数推断规则\n- 用户说\"生成/画/创建\"图片 → `gen`\n- 用户说\"编辑/改/换风格\"图片 → `img2img`\n- 用户说\"融合/合成\"多图 → `seq --refs`\n- 用户有额外要求 → 加到描述里\n\n---\n\n## 限制\n\n- 每个模型每天 18 次，超了自动换下一个\n- 4.0 仅支持文生图\n- 多图融合参考图 base64 后不能超过 ~15MB\n- 组图上限 15 张\n\nFile v2.4.0:_meta.json\n\n{\n  \"ownerId\": \"kn75s192k98dtwr83bft4myd0582ytyg\",\n  \"slug\": \"bytedance-ai-image-gen\",\n  \"version\": \"2.4.0\",\n  \"publishedAt\": 1779906400834\n}\n\nFile v2.4.0:skill-card.md\n\n## Description: <br>\nByteDance AI Image Gen helps an agent call Volcengine Ark Doubao Seedream models to generate and edit images from text prompts, reference images, or multiple images. <br>\n\nThis skill is ready for commercial/non-commercial use. <br>\n\n## Publisher: <br>\n[etmnb](https://clawhub.ai/user/etmnb) <br>\n\n### License/Terms of Use: <br>\nMIT-0 <br>\n\n\n## Use Case: <br>\nExternal users and developers use this skill to have an agent run local Python commands that submit image generation, image editing, sequence generation, and multi-image fusion requests to Volcengine Ark Seedream endpoints. The skill returns generated image file paths and can also report usage and history status. <br>\n\n### Deployment Geography for Use: <br>\nGlobal <br>\n\n## Known Risks and Mitigations: <br>\nRisk: The skill can automatically call Volcengine services and consume quota after trigger phrases. <br>\nMitigation: Install only with credentials intended for this use, monitor usage, and restrict deployment to agents allowed to spend that quota. <br>\nRisk: Prompts and reference images may be uploaded to Volcengine and retained locally in generated files or history. <br>\nMitigation: Avoid confidential prompts or private images unless local history, temp-file behavior, and confirmation behavior are adjusted first. <br>\n\n\n## Reference(s): <br>\n- [ClawHub skill page](https://clawhub.ai/etmnb/bytedance-ai-image-gen) <br>\n- [Volcengine Ark image generation API documentation](https://www.volcengine.com/docs/82379/1541523) <br>\n- [Volcengine Ark console](https://console.volcengine.com/ark) <br>\n\n\n## Skill Output: <br>\n**Output Type(s):** [Shell commands, Files, Markdown, Configuration] <br>\n**Output Format:** [Markdown guidance with shell commands and generated local image file paths] <br>\n**Output Parameters:** [1D] <br>\n**Other Properties Related to Output:** [May create local image files, maintain JSON usage history, and require Volcengine Ark API credentials and model endpoint IDs.] <br>\n\n## Skill Version(s): <br>\n2.4.0 (source: server release evidence; artifact frontmatter and script report 2.3.0) <br>\n\n## Ethical Considerations: <br>\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. <br>\n\nArchive v1.1.0: 3 files, 23371 bytes\n\nFiles: _meta.json (141b), doubao_image_gen.py (71543b), SKILL.md (6839b)\n\nFile v1.1.0:SKILL.md\n\n---\nname: bytedance-ai-image-gen\ndescription: >\n  ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。\n  支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。\n  参与火山协作奖励计划免费使用顶级生图模型。\n  智能调度模型，每模型每日18次，超限拒绝。\n  ⏱️ 生成需30-60秒，触发后请耐心等待。\nsummary: \"ByteDance AI Image Gen — 调用豆包 Seedream 生成/编辑图片，支持文生图、图生图、组图、多图融合，参与火山协作奖励计划免费使用\"\ntags:\n  image-generation: \"2.1.0\"\n  byteimage: \"2.1.0\"\n  seedream: \"2.1.0\"\n  text-to-image: \"2.1.0\"\n  image-to-image: \"2.1.0\"\n  volcengine: \"2.1.0\"\nversion: 2.3.0\ntrigger_patterns:\n  - \"生成图片\"\n  - \"画一张图\"\n  - \"创建图片\"\n  - \"制作图片\"\n  - \"画图\"\n  - \"帮我画\"\n  - \"帮我生成\"\n  - \"我想生成\"\n  - \"搞个图\"\n  - \"来张图\"\n  - \"做张图\"\n  - \"generate image\"\n  - \"create image\"\n  - \"txt2img\"\n  - \"please generate\"\n  - \"please draw\"\n  - \"图生图\"\n  - \"图片编辑\"\n  - \"编辑图片\"\n  - \"改图\"\n  - \"换背景\"\n  - \"换风格\"\n  - \"img2img\"\n  - \"image to image\"\n  - \"edit this image\"\n  - \"edit image\"\n  - \"融合\"\n  - \"合成\"\n  - \"两张图片\"\n  - \"多张图片\"\n  - \"拼合\"\n  - \"合图\"\n  - \"blend\"\n  - \"fuse\"\n  - \"merge\"\n  - \"combine\"\n  - \"文生图\"\n  - \"生图\"\n  - \"图片生成\"\n  - \"图像生成\"\n  - \"生成图像\"\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - python\n      env:\n        - ARK_API_KEY\n        - DOUBAO_MODEL_45_ID\n        - DOUBAO_MODEL_50L_ID\n        - DOUBAO_MODEL_40_ID\n    primaryEnv: ARK_API_KEY\n    envVars:\n      - name: ARK_API_KEY\n        required: true\n        description: 火山方舟 API Key，在 console.volcengine.com/ark → API Key 管理 创建\n      - name: DOUBAO_MODEL_45_ID\n        required: true\n        description: Seedream-4.5 接入点 ID，在 console.volcengine.com/ark → 在线推理 获取\n      - name: DOUBAO_MODEL_50L_ID\n        required: true\n        description: Seedream-5.0-Lite 接入点 ID\n      - name: DOUBAO_MODEL_40_ID\n        required: true\n        description: Seedream-4.0 接入点 ID\n      - name: VOLCENGINE_ACCESS_KEY\n        required: false\n        description: 火山 IAM Access Key，在 console.volcengine.com/iam 创建，用于自动同步用量\n      - name: VOLCENGINE_SECRET_KEY\n        required: false\n        description: 火山 IAM Secret Key，与 ACCESS_KEY 配套\n    emoji: \"🎨\"\n    homepage: https://www.volcengine.com/docs/82379/1541523\n    install:\n      - kind: uv\n        package: requests\n---\n\n# ByteDance AI Image Gen — 豆包 Seedream 图片生成\n\n调用火山方舟 Seedream [API 文档](https://www.volcengine.com/docs/82379/1541523?lang=zh) 生成/编辑图片，自动选择模型并限制用量。参与火山协作奖励计划免费使用顶级生图模型。\n\n> ⚠️ **配置只需一次！** 如果 `.env` 文件已存在且包含 `ARK_API_KEY` 和 3 个模型 ID，说明已配置过，**直接跳到\"调用方式\"执行命令，不要重新配置。**\n\n## 🚀 首次配置\n\n### 1. 获取 API Key\n\n1. 打开 https://console.volcengine.com/ark 注册/登录\n2. 左侧菜单 → **API Key 管理** → 创建 API Key → 复制保存\n\n### 2. 创建模型接入点\n\n在同一个控制台，左侧菜单 → **在线推理** → 创建推理接入点，选以下 3 个模型：\n\n| 环境变量名 | 模型 | 说明 |\n|-----------|------|------|\n| `DOUBAO_MODEL_45_ID` | Seedream-4.5 | 主力模型，所有模式优先 |\n| `DOUBAO_MODEL_50L_ID` | Seedream-5.0-Lite | 轻量备选 |\n| `DOUBAO_MODEL_40_ID` | Seedream-4.0 | 仅支持文生图 |\n\n每个接入点创建后会得到一个 `ep-xxxxx` 格式的 ID，复制保存。\n\n### 3. 配置 .env 文件\n\n在 Skill 目录下创建 `.env` 文件，填入你的 Key 和接入点 ID：\n\n```bash\nARK_API_KEY=你的API Key\nDOUBAO_MODEL_45_ID=ep-xxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxx\n```\n\n> 可选：加上 IAM 密钥可自动同步控制台用量\n> ```bash\n> VOLCENGINE_ACCESS_KEY=你的Access Key\n> VOLCENGINE_SECRET_KEY=你的Secret Key\n> ```\n\n### 4. 测试\n\n```bash\npython doubao_image_gen.py --help\n```\n\n不报错，就配置好了。\n\n---\n\n## ⚡ 调用方式 — 触发后必须立即执行命令\n\n> **这是操作手册，不是参考文档。检测到触发词后，直接复制对应命令执行，禁止只回复文字。**\n\n确认 Skill 目录后，先 `cd` 进去，再执行对应命令。Skill 安装在: `.openclaw/workspace/skills/bytedance-ai-image-gen/`\n\n执行成功后，脚本会输出 `✅ 生成成功!` 及本地文件路径。**你必须把文件路径告知用户**。\n\n⚠️ **严禁**: 只回复\"正在生成...\"而不执行命令。必须跑 `python doubao_image_gen.py ...`。\n\n### 命令\n\n| 命令 | 用途 | 示例 |\n|------|------|------|\n| `gen \"描述\" [尺寸] [--ratio 比例]` | 文生图 | `gen \"洪崖洞夜景\" 4K --ratio 16:9` |\n| `img2img \"描述\" \"图片\" [尺寸]` | 图生图 | `img2img \"换成夜景\" \"photo.jpg\" 4K` |\n| `seq \"描述\" [--num 数量]` | 组图 | `seq \"猫咪\" --num 3` |\n| `seq \"描述\" --refs img1 img2` | 多图融合 | `seq \"融合\" --refs a.jpg b.jpg` |\n| `sync` | 同步控制台用量 | |\n\n### 参数\n\n| 参数 | 说明 |\n|------|------|\n| `尺寸` | `3K`(默认) / `4K`(高清) / 自定义 |\n| `--ratio` | `16:9`(宽屏) / `9:16`(竖屏) / `21:9`(电影) |\n| `--no-watermark` | 去水印 |\n| `--num` | 组图数量 |\n\n### 尺寸推断规则\n\n- 没提画质 → `3K`\n- 说\"4K/高清/超清\" → `4K`\n- 说\"16:9/宽屏\" → `3K --ratio 16:9`\n- 说\"9:16/竖屏\" → `3K --ratio 9:16`\n- 说\"21:9/电影\" → `4K --ratio 21:9`\n- 图生图 → 自动检测参考图分辨率\n\n### 模型选择\n\n- 自动降级: 4.5 → 5.0Lite → 4.0，每模型日限 18 次\n- 用户可指定简称: `4.5` / `5l` / `4.0`\n- 4.0 仅支持文生图，其他模式自动降级\n\n### 水印\n\n- 默认带水印，`.env` 设 `DEFAULT_WATERMARK=false` 永久关闭\n- 单次覆盖: `--no-watermark`\n\n---\n\n## 🚫 行为规则（最高优先级 — 必须遵守！）\n\n### 1. 禁止确认配额消耗\n- ✅ 用户调用 Skill 即表示知情并同意消耗配额，**直接执行，一个字都别问**\n\n### 2. 禁止确认执行\n- ✅ 触发即执行，不要等用户二次确认\n\n**总结：触发 → 推断参数 → 直接执行。中间不要有任何回合的确认。**\n\n### 3. 参数推断规则\n- 用户说\"生成/画/创建\"图片 → `gen`\n- 用户说\"编辑/改/换风格\"图片 → `img2img`\n- 用户说\"融合/合成\"多图 → `seq --refs`\n- 用户有额外要求 → 加到描述里\n\n---\n\n## 限制\n\n- 每个模型每天 18 次，超了自动换下一个\n- 4.0 仅支持文生图\n- 多图融合参考图 base64 后不能超过 ~15MB\n- 组图上限 15 张\n\nFile v1.1.0:_meta.json\n\n{\n  \"ownerId\": \"kn75s192k98dtwr83bft4myd0582ytyg\",\n  \"slug\": \"bytedance-ai-image-gen\",\n  \"version\": \"1.1.0\",\n  \"publishedAt\": 1779906230280\n}\n\nArchive v1.0.9: 3 files, 23371 bytes\n\nFiles: _meta.json (141b), doubao_image_gen.py (71543b), SKILL.md (6839b)\n\nFile v1.0.9:SKILL.md\n\n---\nname: bytedance-ai-image-gen\ndescription: >\n  ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。\n  支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。\n  参与火山协作奖励计划免费使用顶级生图模型。\n  智能调度模型，每模型每日18次，超限拒绝。\n  ⏱️ 生成需30-60秒，触发后请耐心等待。\nsummary: \"ByteDance AI Image Gen — 调用豆包 Seedream 生成/编辑图片，支持文生图、图生图、组图、多图融合，参与火山协作奖励计划免费使用\"\ntags:\n  image-generation: \"2.1.0\"\n  byteimage: \"2.1.0\"\n  seedream: \"2.1.0\"\n  text-to-image: \"2.1.0\"\n  image-to-image: \"2.1.0\"\n  volcengine: \"2.1.0\"\nversion: 2.3.0\ntrigger_patterns:\n  - \"生成图片\"\n  - \"画一张图\"\n  - \"创建图片\"\n  - \"制作图片\"\n  - \"画图\"\n  - \"帮我画\"\n  - \"帮我生成\"\n  - \"我想生成\"\n  - \"搞个图\"\n  - \"来张图\"\n  - \"做张图\"\n  - \"generate image\"\n  - \"create image\"\n  - \"txt2img\"\n  - \"please generate\"\n  - \"please draw\"\n  - \"图生图\"\n  - \"图片编辑\"\n  - \"编辑图片\"\n  - \"改图\"\n  - \"换背景\"\n  - \"换风格\"\n  - \"img2img\"\n  - \"image to image\"\n  - \"edit this image\"\n  - \"edit image\"\n  - \"融合\"\n  - \"合成\"\n  - \"两张图片\"\n  - \"多张图片\"\n  - \"拼合\"\n  - \"合图\"\n  - \"blend\"\n  - \"fuse\"\n  - \"merge\"\n  - \"combine\"\n  - \"文生图\"\n  - \"生图\"\n  - \"图片生成\"\n  - \"图像生成\"\n  - \"生成图像\"\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - python\n      env:\n        - ARK_API_KEY\n        - DOUBAO_MODEL_45_ID\n        - DOUBAO_MODEL_50L_ID\n        - DOUBAO_MODEL_40_ID\n    primaryEnv: ARK_API_KEY\n    envVars:\n      - name: ARK_API_KEY\n        required: true\n        description: 火山方舟 API Key，在 console.volcengine.com/ark → API Key 管理 创建\n      - name: DOUBAO_MODEL_45_ID\n        required: true\n        description: Seedream-4.5 接入点 ID，在 console.volcengine.com/ark → 在线推理 获取\n      - name: DOUBAO_MODEL_50L_ID\n        required: true\n        description: Seedream-5.0-Lite 接入点 ID\n      - name: DOUBAO_MODEL_40_ID\n        required: true\n        description: Seedream-4.0 接入点 ID\n      - name: VOLCENGINE_ACCESS_KEY\n        required: false\n        description: 火山 IAM Access Key，在 console.volcengine.com/iam 创建，用于自动同步用量\n      - name: VOLCENGINE_SECRET_KEY\n        required: false\n        description: 火山 IAM Secret Key，与 ACCESS_KEY 配套\n    emoji: \"🎨\"\n    homepage: https://www.volcengine.com/docs/82379/1541523\n    install:\n      - kind: uv\n        package: requests\n---\n\n# ByteDance AI Image Gen — 豆包 Seedream 图片生成\n\n调用火山方舟 Seedream [API 文档](https://www.volcengine.com/docs/82379/1541523?lang=zh) 生成/编辑图片，自动选择模型并限制用量。参与火山协作奖励计划免费使用顶级生图模型。\n\n> ⚠️ **配置只需一次！** 如果 `.env` 文件已存在且包含 `ARK_API_KEY` 和 3 个模型 ID，说明已配置过，**直接跳到\"调用方式\"执行命令，不要重新配置。**\n\n## 🚀 首次配置\n\n### 1. 获取 API Key\n\n1. 打开 https://console.volcengine.com/ark 注册/登录\n2. 左侧菜单 → **API Key 管理** → 创建 API Key → 复制保存\n\n### 2. 创建模型接入点\n\n在同一个控制台，左侧菜单 → **在线推理** → 创建推理接入点，选以下 3 个模型：\n\n| 环境变量名 | 模型 | 说明 |\n|-----------|------|------|\n| `DOUBAO_MODEL_45_ID` | Seedream-4.5 | 主力模型，所有模式优先 |\n| `DOUBAO_MODEL_50L_ID` | Seedream-5.0-Lite | 轻量备选 |\n| `DOUBAO_MODEL_40_ID` | Seedream-4.0 | 仅支持文生图 |\n\n每个接入点创建后会得到一个 `ep-xxxxx` 格式的 ID，复制保存。\n\n### 3. 配置 .env 文件\n\n在 Skill 目录下创建 `.env` 文件，填入你的 Key 和接入点 ID：\n\n```bash\nARK_API_KEY=你的API Key\nDOUBAO_MODEL_45_ID=ep-xxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxx\n```\n\n> 可选：加上 IAM 密钥可自动同步控制台用量\n> ```bash\n> VOLCENGINE_ACCESS_KEY=你的Access Key\n> VOLCENGINE_SECRET_KEY=你的Secret Key\n> ```\n\n### 4. 测试\n\n```bash\npython doubao_image_gen.py --help\n```\n\n不报错，就配置好了。\n\n---\n\n## ⚡ 调用方式 — 触发后必须立即执行命令\n\n> **这是操作手册，不是参考文档。检测到触发词后，直接复制对应命令执行，禁止只回复文字。**\n\n确认 Skill 目录后，先 `cd` 进去，再执行对应命令。Skill 安装在: `.openclaw/workspace/skills/bytedance-ai-image-gen/`\n\n执行成功后，脚本会输出 `✅ 生成成功!` 及本地文件路径。**你必须把文件路径告知用户**。\n\n⚠️ **严禁**: 只回复\"正在生成...\"而不执行命令。必须跑 `python doubao_image_gen.py ...`。\n\n### 命令\n\n| 命令 | 用途 | 示例 |\n|------|------|------|\n| `gen \"描述\" [尺寸] [--ratio 比例]` | 文生图 | `gen \"洪崖洞夜景\" 4K --ratio 16:9` |\n| `img2img \"描述\" \"图片\" [尺寸]` | 图生图 | `img2img \"换成夜景\" \"photo.jpg\" 4K` |\n| `seq \"描述\" [--num 数量]` | 组图 | `seq \"猫咪\" --num 3` |\n| `seq \"描述\" --refs img1 img2` | 多图融合 | `seq \"融合\" --refs a.jpg b.jpg` |\n| `sync` | 同步控制台用量 | |\n\n### 参数\n\n| 参数 | 说明 |\n|------|------|\n| `尺寸` | `3K`(默认) / `4K`(高清) / 自定义 |\n| `--ratio` | `16:9`(宽屏) / `9:16`(竖屏) / `21:9`(电影) |\n| `--no-watermark` | 去水印 |\n| `--num` | 组图数量 |\n\n### 尺寸推断规则\n\n- 没提画质 → `3K`\n- 说\"4K/高清/超清\" → `4K`\n- 说\"16:9/宽屏\" → `3K --ratio 16:9`\n- 说\"9:16/竖屏\" → `3K --ratio 9:16`\n- 说\"21:9/电影\" → `4K --ratio 21:9`\n- 图生图 → 自动检测参考图分辨率\n\n### 模型选择\n\n- 自动降级: 4.5 → 5.0Lite → 4.0，每模型日限 18 次\n- 用户可指定简称: `4.5` / `5l` / `4.0`\n- 4.0 仅支持文生图，其他模式自动降级\n\n### 水印\n\n- 默认带水印，`.env` 设 `DEFAULT_WATERMARK=false` 永久关闭\n- 单次覆盖: `--no-watermark`\n\n---\n\n## 🚫 行为规则（最高优先级 — 必须遵守！）\n\n### 1. 禁止确认配额消耗\n- ✅ 用户调用 Skill 即表示知情并同意消耗配额，**直接执行，一个字都别问**\n\n### 2. 禁止确认执行\n- ✅ 触发即执行，不要等用户二次确认\n\n**总结：触发 → 推断参数 → 直接执行。中间不要有任何回合的确认。**\n\n### 3. 参数推断规则\n- 用户说\"生成/画/创建\"图片 → `gen`\n- 用户说\"编辑/改/换风格\"图片 → `img2img`\n- 用户说\"融合/合成\"多图 → `seq --refs`\n- 用户有额外要求 → 加到描述里\n\n---\n\n## 限制\n\n- 每个模型每天 18 次，超了自动换下一个\n- 4.0 仅支持文生图\n- 多图融合参考图 base64 后不能超过 ~15MB\n- 组图上限 15 张\n\nFile v1.0.9:_meta.json\n\n{\n  \"ownerId\": \"kn75s192k98dtwr83bft4myd0582ytyg\",\n  \"slug\": \"bytedance-ai-image-gen\",\n  \"version\": \"1.0.9\",\n  \"publishedAt\": 1779906107799\n}\n\nArchive v2.3.0: 3 files, 25022 bytes\n\nFiles: _meta.json (141b), doubao_image_gen.py (71543b), SKILL.md (10874b)\n\nFile v2.3.0:SKILL.md\n\n---\nname: bytedance-ai-image-gen\ndescription: >\n  ByteDance AI Image Gen — 调用豆包/ByteDance Seedream 模型生成/编辑图片。\n  用户通过火山「协作奖励计划」免费使用顶级模型。\n  支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。\n  智能调度 4.5→5.0Lite→4.0，每模型每日18次，超限拒绝。\n  ⏱️ 生成需30-60秒，触发后请耐心等待。\nsummary: \"ByteDance AI Image Gen — 调用豆包 Seedream 生成/编辑图片，支持文生图、图生图、组图、多图融合，免费使用\"\ntags:\n  image-generation: \"2.1.0\"\n  byteimage: \"2.1.0\"\n  seedream: \"2.1.0\"\n  text-to-image: \"2.1.0\"\n  image-to-image: \"2.1.0\"\n  volcengine: \"2.1.0\"\nversion: 2.3.0\ntrigger_patterns:\n  - \"生成图片\"\n  - \"画一张图\"\n  - \"创建图片\"\n  - \"制作图片\"\n  - \"画图\"\n  - \"帮我画\"\n  - \"帮我生成\"\n  - \"我想生成\"\n  - \"搞个图\"\n  - \"来张图\"\n  - \"做张图\"\n  - \"generate image\"\n  - \"create image\"\n  - \"txt2img\"\n  - \"please generate\"\n  - \"please draw\"\n  - \"图生图\"\n  - \"图片编辑\"\n  - \"编辑图片\"\n  - \"改图\"\n  - \"换背景\"\n  - \"换风格\"\n  - \"img2img\"\n  - \"image to image\"\n  - \"edit this image\"\n  - \"edit image\"\n  - \"融合\"\n  - \"合成\"\n  - \"两张图片\"\n  - \"多张图片\"\n  - \"拼合\"\n  - \"合图\"\n  - \"blend\"\n  - \"fuse\"\n  - \"merge\"\n  - \"combine\"\n  - \"文生图\"\n  - \"生图\"\n  - \"图片生成\"\n  - \"图像生成\"\n  - \"生成图像\"\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - python\n      env:\n        - ARK_API_KEY\n        - DOUBAO_MODEL_45_ID\n        - DOUBAO_MODEL_50L_ID\n        - DOUBAO_MODEL_40_ID\n    primaryEnv: ARK_API_KEY\n    envVars:\n      - name: ARK_API_KEY\n        required: true\n        description: 火山方舟 API Key，在 console.volcengine.com/ark → API Key 管理 创建\n      - name: DOUBAO_MODEL_45_ID\n        required: true\n        description: Seedream-4.5 接入点 ID，在 console.volcengine.com/ark → 在线推理 获取\n      - name: DOUBAO_MODEL_50L_ID\n        required: true\n        description: Seedream-5.0-Lite 接入点 ID\n      - name: DOUBAO_MODEL_40_ID\n        required: true\n        description: Seedream-4.0 接入点 ID\n      - name: VOLCENGINE_ACCESS_KEY\n        required: false\n        description: 火山 IAM Access Key，在 console.volcengine.com/iam 创建，用于自动同步用量\n      - name: VOLCENGINE_SECRET_KEY\n        required: false\n        description: 火山 IAM Secret Key，与 ACCESS_KEY 配套\n    emoji: \"🎨\"\n    homepage: https://www.volcengine.com/docs/82379/1541523\n    install:\n      - kind: uv\n        package: requests\n---\n\n# ByteDance AI Image Gen — 豆包/ByteDance Seedream 图片生成\n\n调用火山方舟 Seedream [API 文档](https://www.volcengine.com/docs/82379/1541523?lang=zh) 生成/编辑图片，自动选择模型并限制用量。\n\n> ⏱️ 图片生成通常需要 **30-60 秒**，其间无额外反馈。触发后请耐心等待，不要重复发送。\n\n## ⚡ 调用方式 — 触发后必须立即执行命令\n\n> **这是操作手册，不是参考文档。检测到触发词后，直接复制对应命令执行，禁止只回复文字。**\n\n确认 Skill 目录后，先 `cd` 进去，再执行对应命令。Skill 安装在: `.openclaw/workspace/skills/bytedance-ai-image-gen/`\n\n执行成功后，脚本会输出 `✅ 生成成功!` 及本地文件路径。**你必须把以下信息告知用户**:\n- 模型、分辨率、比例\n- 本地文件路径（`文件: .../Image/xxx.png`）← 这行一定要报\n- 可以不报临时 CDN URL（会过期）\n\n### 文生图\n```bash\npython doubao_image_gen.py gen \"用户描述\" [尺寸] [--ratio 比例] [--no-watermark]\n```\n示例: `python doubao_image_gen.py gen \"洪崖洞夜景\" 4K --ratio 16:9`\n\n### 图生图\n```bash\npython doubao_image_gen.py img2img \"用户描述\" \"图片路径或URL\" [尺寸] [--no-watermark]\n```\n示例: `python doubao_image_gen.py img2img \"换成夜景风格\" \"photo.jpg\" 4K`\n\n### 组图（一次多张）\n```bash\npython doubao_image_gen.py seq \"用户描述\" [--num 数量] [尺寸] [--no-watermark]\n```\n\n### 多图融合\n```bash\npython doubao_image_gen.py seq \"融合描述\" --refs img1.jpg img2.jpg [--no-watermark]\n```\n\n⚠️ **严禁**: 只回复\"正在生成...\"而不执行命令。必须跑 `python doubao_image_gen.py ...`。\n\n---\n\n## 功能\n\n| 模式 | 说明 | 支持 |\n|------|------|------|\n| 文生单图 | 纯文本描述生成 1 张图片 | ✅ |\n| 文生组图 | 纯文本描述一次生成多张 | ✅ |\n| 图生单图 | 基于参考图编辑/风格转换 | ✅ |\n| 图生组图 | 基于参考图一次生成多张 | ✅ |\n| 多图融合 | 多张参考图融合为新图 | ✅ |\n| 用量查询 | 控制台用量同步（`python doubao_image_gen.py sync`） | ✅ |\n\n---\n\n## 🚫 行为规则（最高优先级 — 必须遵守！）\n\n**以下行为严格禁止，违反任一即为 Bug：**\n\n### 0. 默认 CLI 模式\n- ✅ 默认使用 CLI 命令：`python doubao_image_gen.py <命令> <参数...>`\n- 也可用 Python SDK 调用，但不推荐\n\n### 1. 禁止确认配额消耗\n- ✅ 用户调用 Skill 即表示知情并同意消耗配额，**直接执行，一个字都别问**\n\n### 尺寸自适应\n- ✅ 自动推断并直接执行：\n  - 没提画质 → `size=\"3K\"`（不传 ratio，API 自适应）\n  - 说\"4K/高清/超清\" → `size=\"4K\"`（不传 ratio）\n  - 说\"4K+16:9\" → `size=\"4K\", ratio=\"16:9\"`\n  - 说\"16:9/宽屏\" → `size=\"3K\", ratio=\"16:9\"`\n  - 说\"9:16/竖屏\" → `size=\"3K\", ratio=\"9:16\"`\n  - 说\"21:9/电影\" → `size=\"4K\", ratio=\"21:9\"`\n  - 明确写了尺寸 → 按指定\n- 图生图 → 自动检测参考图分辨率\n\n### 2.5 模型选择\n- 默认按优先级自动降级: 4.5 → 5.0L → 4.0\n- 用户可指定模型。简称映射：\n  - `4.5` / `4-5` / `seedream4.5` → `Doubao-Seedream-4.5`\n  - `5l` / `5.0l` / `5lite` / `seedream5lite` → `Doubao-Seedream-5.0-Lite`\n  - `4.0` / `40` / `seedream4` → `Doubao-Seedream-4.0`\n- AI 根据映射补全完整名称，传给脚本 `force_model` 参数\n- 每模型日限 18 次，超限自动跳过\n- ✅ 不询问用哪个模型，除非用户明确指定\n\n### 3. 禁止确认执行\n- ✅ 触发即执行，不要等用户二次确认\n\n**总结：触发 → 推断参数 → 直接执行。中间不要有任何回合的确认。**\n\n### 4. 水印控制\n- 默认带水印，`.env` 设 `DEFAULT_WATERMARK=false` 永久关闭\n- 单次覆盖：`--no-watermark` 强制关，不传则用默认\n\n---\n\n## 触发词\n\n### 文生图触发\n\n| 中文 | 英文 |\n|------|------|\n| `生成图片` `画一张图` `创建图片` `制作图片` | `generate image` `create image` `txt2img` |\n| `画图` `帮我画` `帮我生成` `我想生成` | `please generate` `please draw` `I want to generate` |\n| `搞个图` `来张图` `做张图` `出个图` | `draw` `make an image` `create a picture` |\n\n### 图生图 / 图片编辑触发\n\n| 中文 | 英文 |\n|------|------|\n| `图生图` `图片编辑` `编辑图片` `改图` | `img2img` `image to image` `edit this image` |\n| `换背景` `换风格` `风格转换` `图片处理` | `edit image` `modify image` `transform image` |\n\n### 多图融合 / 合成触发\n\n| 中文 | 英文 |\n|------|------|\n| `融合` `合成` `两张图片` `多张图片` | `blend` `fuse` `merge` `combine` |\n| `拼合` `合图` `融图` `图片混合` | `composite` `mix images` |\n\n### 带风格前缀\n\n`极致生成:` `快速草稿:` `艺术风格:` `4K高清:`\n\n## 工作流程\n\n调用 `doubao_image_gen.py` 脚本：\n\n1. **检测触发词** — 从用户输入提取 prompt、size、style、图生图标记\n2. **模型选择** — 按 4.5 → 5.0Lite → 4.0 降级，满 18 次跳过\n3. **尺寸自动推断** — 不再询问用户，直接按规则确定\n4. **图片规范化** — 本地路径自动 base64 编码，URL 原样通过\n5. **调用 API** — 请求火山方舟 Seedream 图片生成接口\n6. **保存图片** — 下载到 `Image/`，命名：`内容分辨率-模型-时间.png`\n7. **记录历史** — 写入 `image_gen_history.json`，7 天自动清理\n\n## 命令参考\n\n运行 `python doubao_image_gen.py --help` 查看完整参数列表和调用示例。\n\n## 模型\n\n| 模型 | 优先级 | 日限 | 文生图 | 图生图 | 组图/融合 |\n|------|--------|------|--------|--------|-----------|\n| Seedream-4.5 | 1 | 18 | ✅ | ✅ | ✅ |\n| Seedream-5.0-Lite | 2 | 18 | ✅ | ✅ | ✅ |\n| Seedream-4.0 | 3 | 18 | ✅ | ❌ 自动降级 | ❌ 自动降级 |\n\n> **已知限制**: 4.0 仅支持文生单图，其他模式自动降级到 4.5/5.0L。多图融合时注意参考图累计大小，base64 后超过 ~15MB 会被 API 拒绝。组图上限为 15 张 (参考图 + 输出)。\n\n超限返回 `{\"success\":false,\"error\":\"...\"}`。\n\n## 🎁 火山「协作奖励计划」\n\n注册火山方舟即可**免费使用** Seedream-4.5、5.0-Lite、4.0 等顶级生图模型，**无需预充值**。详情见 [火山方舟文档](https://www.volcengine.com/docs/82379/1391869)。\n\n## 配置（.env 文件）\n\n在 Skill 目录下创建 `.env` 文件：\n\n```bash\n# 必填\nARK_API_KEY=ark-your-key-here\nDOUBAO_MODEL_45_ID=ep-xxxxxxxxxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxxxxxxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxxxxxxxxx\n\n# 可选\nVOLCENGINE_ACCESS_KEY=AKLTxxx          # 控制台用量同步\nVOLCENGINE_SECRET_KEY=your-secret       # 控制台用量同步\n```\n\n> MODEL_ID 既是 API 调用 ID 也是控制台端点 ID，一个变量两用。\n\n脚本启动时自动加载 `.env`，无需手动 export。\n\n### 获取密钥\n\n- **ARK_API_KEY**: https://console.volcengine.com/ark → API Key 管理\n- **MODEL_ID**: https://console.volcengine.com/ark → 在线推理 → 点击接入点名称获取端点ID\n- **IAM 密钥**: https://console.volcengine.com/iam → 访问控制 → 密钥管理\n\n## 文件\n\n| 文件 | 说明 |\n|------|------|\n| `doubao_image_gen.py` | 核心脚本 |\n| `.env` | 环境变量配置 |\n| `image_gen_history.json` | 调用历史+计数 (7天清理) |\n| `Image/` | 生成图片保存目录 |\n| `Temp/` | 图生图本地缓存（自动创建/清理） |\n\n---\n\n## 常见错误\n\n| 错误码 | 原因 | 解决 |\n|--------|------|------|\n| `HTTP 429` | 每分钟调用频率超限 | 等待 1-2 分钟后重试 |\n| `HTTP 400` | 参数错误（尺寸/模型不支持） | 检查参数格式 |\n| `InvalidEndpointOrModel.NotFound` | 模型 ID 无效 | 检查 .env 中的 MODEL_ID |\n| `所有模型调用次数已达上限` | 三个模型都满 18 次 | 等明天重置，或查控制台确认 |\n| `图片文件不存在` | 图生图参考图路径有误 | 检查文件路径，确保文件存在 |\n| `IAM 密钥未配置` | 缺少 VOLCENGINE_ACCESS_KEY | .env 中配置 IAM 密钥 |\n\n## 💬 支持与反馈\n\n如果遇到错误或问题，请在 [clawhub.ai 技能页面](https://clawhub.ai/etmnb/bytedance-ai-image-gen) 下方留言反馈。\n\nFile v2.3.0:_meta.json\n\n{\n  \"ownerId\": \"kn75s192k98dtwr83bft4myd0582ytyg\",\n  \"slug\": \"bytedance-ai-image-gen\",\n  \"version\": \"2.3.0\",\n  \"publishedAt\": 1779728586002\n}","readmeExcerpt":"Skill: Bytedance Ai Image Gen Owner: etmnb Summary: ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。 支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。 参与火山协作奖励计划免费使用顶级生图模型。 智能调度模型，每模型每日18次，超限拒绝。 ⏱️ 生成需30-60秒，触发后请耐心等待。 Tags: latest:2.4.2 Version history: v2.4.2 | 2026-05-28T03:04:09.927Z | auto bytedance-ai-image-gen v2.4.2 - No code or functionality changes. - Internal metadata updated only. - No impact on user experience or feature","codeSnippets":[],"executableExamples":[{"language":"bash","snippet":"ARK_API_KEY=你的API Key\nDOUBAO_MODEL_45_ID=ep-xxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxx"},{"language":"bash","snippet":"> VOLCENGINE_ACCESS_KEY=你的Access Key\n> VOLCENGINE_SECRET_KEY=你的Secret Key\n>"},{"language":"bash","snippet":"python doubao_image_gen.py --help"},{"language":"bash","snippet":"ARK_API_KEY=你的API Key\nDOUBAO_MODEL_45_ID=ep-xxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxx"},{"language":"bash","snippet":"> VOLCENGINE_ACCESS_KEY=你的Access Key\n> VOLCENGINE_SECRET_KEY=你的Secret Key\n>"},{"language":"bash","snippet":"python doubao_image_gen.py --help"}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: bytedance-ai-image-gen\ndescription: >\n  ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。\n  支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。\n  参与火山协作奖励计划免费使用顶级生图模型。\n  智能调度模型，每模型每日18次，超限拒绝。\n  ⏱️ 生成需30-60秒，触发后请耐心等待。\nsummary: \"ByteDance AI Image Gen — 调用豆包 Seedream 生成/编辑图片，支持文生图、图生图、组图、多图融合，参与火山协作奖励计划免费使用\"\ntags:\n  image-generation: \"2.1.0\"\n  byteimage: \"2.1.0\"\n  seedream: \"2.1.0\"\n  text-to-image: \"2.1.0\"\n  image-to-image: \"2.1.0\"\n  volcengine: \"2.1.0\"\nversion: 2.3.0\ntrigger_patterns:\n  - \"生成图片\"\n  - \"画一张图\"\n  - \"创建图片\"\n  - \"制作图片\"\n  - \"画图\"\n  - \"帮我画\"\n  - \"帮我生成\"\n  - \"我想生成\"\n  - \"搞个图\"\n  - \"来张图\"\n  - \"做张图\"\n  - \"generate image\"\n  - \"create image\"\n  - \"txt2img\"\n  - \"please generate\"\n  - \"please draw\"\n  - \"图生图\"\n  - \"图片编辑\"\n  - \"编辑图片\"\n  - \"改图\"\n  - \"换背景\"\n  - \"换风格\"\n  - \"img2img\"\n  - \"image to image\"\n  - \"edit this image\"\n  - \"edit image\"\n  - \"融合\"\n  - \"合成\"\n  - \"两张图片\"\n  - \"多张图片\"\n  - \"拼合\"\n  - \"合图\"\n  - \"blend\"\n  - \"fuse\"\n  - \"merge\"\n  - \"combine\"\n  - \"文生图\"\n  - \"生图\"\n  - \"图片生成\"\n  - \"图像生成\"\n  - \"生成图像\"\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - python\n      env:\n        - ARK_API_KEY\n        - DOUBAO_MODEL_45_ID\n        - DOUBAO_MODEL_50L_ID\n        - DOUBAO_MODEL_40_ID\n    primaryEnv: ARK_API_KEY\n    envVars:\n      - name: ARK_API_KEY\n        required: true\n        description: 火山方舟 API Key，在 console.volcengine.com/ark → API Key 管理 创建\n      - name: DOUBAO_MODEL_45_ID\n        required: true\n        description: Seedream-4.5 接入点 ID，在 console.volcengine.com/ark → 在线推理 获取\n      - name: DOUBAO_MODEL_50L_ID\n        required: true\n        description: Seedream-5.0-Lite 接入点 ID\n      - name: DOUBAO_MODEL_40_ID\n        required: true\n        description: Seedream-4.0 接入点 ID\n      - name: VOLCENGINE_ACCESS_KEY\n        required: false\n        description: 火山 IAM Access Key，在 console.volcengine.com/iam 创建，用于自动同步用量\n      - name: VOLCENGINE_SECRET_KEY\n        required: false\n        description: 火山 IAM Secret Key，与 ACCESS_KEY 配套\n    emoji: \"🎨\"\n    homepage: https://www.volcengine.com/docs/82379/1541523\n    install:\n      - kind: uv\n        package: requests\n---\n\n# ByteDance AI Image Gen — 豆包 Seedream 图片生成\n\n调用火山方舟 Seedream [API 文档](https://www.volcengine.com/docs/82379/1541523?lang=zh) 生成/编辑图片，自动选择模型并限制用量。参与火山协作奖励计划免费使用顶级生图模型。\n\n> ⚠️ **配置只需一次！** 如果 `.env` 文件已存在且包含 `ARK_API_KEY` 和 3 个模型 ID，说明已配置过，**直接跳到\"调用方式\"执行命令，不要重新配置。**\n\n## 🚀 首次配置\n\n### 1. 获取 API Key\n\n1. 打开 https://console.volcengine.com/ark 注册/登录\n2. 左侧菜单 → **API Key 管理** → 创建 API Key → 复制保存\n\n### 2. 创建模型接入点\n\n在同一个控制台，左侧菜单 → **在线推理** → 创建推理接入点，选以下 3 个模型：\n\n| 环境变量名 | 模型 | 说明 |\n|-----------|------|------|\n| `DOUBAO_MODEL_45_ID` | Seedream-4.5 | 主力模型，所有模式优先 |\n| `DOUBAO_MODEL_50L_ID` | Seedream-5.0-Lite | 轻量备选 |\n| `DOUBAO_MODEL_40_ID` | Seedream-4.0 | 仅支持文生图 |\n\n每个接入点创建后会得到一个 `ep-xxxxx` 格式的 ID，复制保存。\n\n### 3. 配置 .env 文件\n\n在 Skill 目录下创建 `.env` 文件，填入你的 Key 和接入点 ID：\n\n```bash\nARK_API_KEY=你的API Key\nDOUBAO_MODEL_45_ID=ep-xxxxx\nDOUBAO_MODEL_50L_ID=ep-xxxxx\nDOUBAO_MODEL_40_ID=ep-xxxxx\n```\n\n> 可选：加上 IAM 密钥可自动同步控制台用量\n> ```bash\n> VOLCENGINE_ACCESS_KEY"},{"path":"_meta.json","content":"{\n  \"ownerId\": \"kn75s192k98dtwr83bft4myd0582ytyg\",\n  \"slug\": \"bytedance-ai-image-gen\",\n  \"version\": \"2.4.2\",\n  \"publishedAt\": 1779937449927\n}"},{"path":"skill-card.md","content":"## Description:\n\nByteDance AI Image Gen calls Volcengine Ark Doubao Seedream models to generate and edit single images, image sets, image-to-image edits, and multi-image fusions with automatic model selection and daily usage limits.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[etmnb](https://clawhub.ai/user/etmnb)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nDevelopers and external users use this skill to generate or edit images with Volcengine Ark Doubao Seedream models from natural-language prompts, reference images, or multiple images. It helps agents choose model and size parameters, run the Python CLI, and return generated local image paths.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: Prompts and reference images may be sent to Volcengine services.\n\nMitigation: Use only prompts and images that are acceptable to share with Volcengine; avoid sensitive local images.\n\nRisk: The skill can store prompts, image URLs, usage counts, generated files, and temporary copies locally.\n\nMitigation: Review and periodically clear the local history, Image, and Temp directories, and set storage paths deliberately when needed.\n\nRisk: The skill instructs agents to execute generation without a second confirmation, which can consume service quota.\n\nMitigation: Review trigger behavior before installation and restrict use to contexts where automatic execution is acceptable.\n\nRisk: API keys and model endpoint identifiers are required for operation.\n\nMitigation: Store secrets in secure environment management rather than a skill-directory .env file when possible.\n\n## Reference(s):\n\n- [ClawHub Skill Page](https://clawhub.ai/etmnb/skills/bytedance-ai-image-gen)\n- [Volcengine Ark Seedream API Documentation](https://www.volcengine.com/docs/82379/1541523)\n- [Volcengine Ark Console](https://console.volcengine.com/ark)\n\n## Skill Output:\n\n**Output Type(s):** [Shell commands, Files, Guidance]\n\n**Output Format:** [Console text with generated image file paths and status details]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [May save generated images under a local Image directory and keep seven-day prompt and usage history.]\n\n## Skill Version(s):\n\n2.4.2 (source: server release metadata; artifact frontmatter reports 2.3.0 and artifact _meta.json reports 2.3.1)\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":"ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。 支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。 参与火山协作奖励计划免费使用顶级生图模型。 智能调度模型，每模型每日18次，超限拒绝。 ⏱️ 生成需30-60秒，触发后请耐心等待。 Skill: Bytedance Ai Image Gen Owner: etmnb Summary: ByteDance AI Image Gen — 调用豆包 Seedream 模型生成/编辑图片。 支持文生单图、文生组图、图生单图、图生组图、多图融合、图片编辑。 参与火山协作奖励计划免费使用顶级生图模型。 智能调度模型，每模型每日18次，超限拒绝。 ⏱️ 生成需30-60秒，触发后请耐心等待。 Tags: latest:2.4.2 Version history: v2.4.2 | 2026-05-28T03:04:09.927Z | auto bytedance-ai-image-gen v2.4.2 - No code or functionality changes. - Internal metadata updated only. - No impact on user experience or feature","editorialQuality":{"score":100,"threshold":65,"status":"ready","wordCount":1278,"uniquenessScore":46,"reasons":[]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-10T04:06:12.435Z","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":"2026-10-10T04:06:12.435Z","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":"2026-10-10T07:42:13.209Z","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"}]}}}