{"slug":"video-understanding","title":"\u89c6\u9891\u7ed3\u6784\u5316\u7406\u89e3","description":"\u628a\u89c6\u9891\u62c6\u6210\u7ed3\u6784\u5316\u7d22\u5f15\uff1a\u573a\u666f\u68c0\u6d4b\u3001ASR \u8f6c\u5199\u3001\u9010\u573a\u666f\u89c2\u5bdf\u3001\u9759\u97f3\u7a97\u53e3\u4e0e\u5199\u4f5c\u7b80\u62a5\u3002","category":"\u9009\u9898\u4e0e\u7814\u7a76","tags":["\u89c6\u9891","\u5206\u6790","\u9009\u9898"],"author":"PiteChen","source_url":"https://github.com/zenstory-ai/video-recap-skills/blob/main/skills/video-understanding/SKILL.md","license":"MIT","status":"published","verification":"source_reviewed","verification_notes":"\u6765\u6e90 zenstory-ai/video-recap-skills\uff08skills/video-understanding/SKILL.md\uff09\uff0c\u672c\u5730\u6253\u5305\u4e0a\u4f20\uff1b\u672a\u6267\u884c\u5305\u5185\u811a\u672c\u3002","revision":"91b4384f15264b298d5723bfc39b9bb8","created_at":"2026-10-28T08:54:28+00:00","updated_at":"2026-10-28T08:54:50+00:00","source_repo":"","source_path":"","source_commit":"","source_updated_at":"","stars":0,"proxy_score":null,"score":null,"grade":null,"dimensions":{},"evaluation":{},"installs":0,"platforms":[],"formats":[],"category_slug":"research","category_name":"\u9009\u9898\u4e0e\u7814\u7a76","usage_guide":{},"read_url":"https://shiliuai.cn/api/v2/skills/video-understanding","skill_url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/SKILL.md","download_url":"https://shiliuai.cn/api/v2/skills/video-understanding/download","install_url":"https://shiliuai.cn/api/v2/skills/video-understanding/install","files":[{"path":"SKILL.md","bytes":8117,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/SKILL.md"},{"path":"references/data-schema.md","bytes":6218,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/references/data-schema.md"},{"path":"references/prompt-templates.md","bytes":1708,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/references/prompt-templates.md"},{"path":"references/research-guide.md","bytes":2880,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/references/research-guide.md"},{"path":"scripts/agent_text.py","bytes":7379,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/agent_text.py"},{"path":"scripts/asr.py","bytes":11581,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/asr.py"},{"path":"scripts/asr_timing_evidence.py","bytes":9958,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/asr_timing_evidence.py"},{"path":"scripts/briefing/__init__.py","bytes":101,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/briefing/__init__.py"},{"path":"scripts/briefing/builder.py","bytes":28757,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/briefing/builder.py"},{"path":"scripts/briefing/context.py","bytes":15103,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/briefing/context.py"},{"path":"scripts/briefing/inputs.py","bytes":9105,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/briefing/inputs.py"},{"path":"scripts/briefing/timeline.py","bytes":18812,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/briefing/timeline.py"},{"path":"scripts/consolidate.py","bytes":27145,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/consolidate.py"},{"path":"scripts/detect.py","bytes":25952,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/detect.py"},{"path":"scripts/extract.py","bytes":2559,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/extract.py"},{"path":"scripts/index_normalize.py","bytes":14305,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/index_normalize.py"},{"path":"scripts/lib.py","bytes":23644,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/lib.py"},{"path":"scripts/storyboard.py","bytes":24654,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/storyboard.py"},{"path":"scripts/timeline_fusion.py","bytes":3102,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/timeline_fusion.py"},{"path":"scripts/understand.py","bytes":192,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/understand.py"},{"path":"scripts/understanding_brief.py","bytes":8100,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/understanding_brief.py"},{"path":"scripts/understanding_cache.py","bytes":12607,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/understanding_cache.py"},{"path":"scripts/understanding_runner.py","bytes":14528,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/understanding_runner.py"},{"path":"scripts/understanding_storyboard.py","bytes":13614,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/understanding_storyboard.py"},{"path":"scripts/vlm.py","bytes":29116,"url":"https://shiliuai.cn/api/v2/skills/video-understanding/files/scripts/vlm.py"}],"content":"---\nname: video-understanding\nuser-invocable: false\ndescription: \u628a\u89c6\u9891\u5206\u6790\u4e3a\u7ed3\u6784\u5316\u7406\u89e3\u7d22\u5f15\uff1a\u573a\u666f\u68c0\u6d4b\u3001ASR \u8f6c\u5199\u3001\u9010\u573a\u666f VLM \u89c2\u5bdf\u3001\u9759\u97f3\u7a97\u53e3\u3001\u878d\u5408\u65f6\u95f4\u7ebf\u548c\u5199\u4f5c brief\u3002 \u7528\u4e8e\u7406\u89e3\u3001\u7d22\u5f15\u6216\u603b\u7ed3\u89c6\u9891\uff0c\u4e5f\u4f5c\u4e3a\u540e\u7eed\u521b\u4f5c\u524d\u7684\u5206\u6790\u9636\u6bb5\u3002\u8f93\u5165\u89c6\u9891\u6587\u4ef6\uff1b\u8f93\u51fa scenes.json\u3001 asr_result.json\u3001vlm_analysis.json\u3001silence_periods.json\u3001timeline_fusion.json\u3001agent_narration_brief.md\u3002 \u89e6\u53d1\u8bcd\uff1a\u89c6\u9891\u7406\u89e3\u3001\u89c6\u9891\u5206\u6790\u3001\u89c6\u9891\u7d22\u5f15\u3001video understanding\u3001analyze video\u3001\u770b\u61c2\u89c6\u9891\u3002\n---\n\n## 1. \u5b9a\u4f4d\n\n\u672c\u6280\u80fd\u628a\u6e90\u89c6\u9891\u8f6c\u6210 Agent \u4e0e\u4e0b\u6e38\u9636\u6bb5\u53ef\u8bfb\u53d6\u7684\u7406\u89e3\u7d22\u5f15\uff0c\u4ee5**\u7d20\u6750\u89c2\u5bdf\u5458 / \u573a\u8bb0**\u7684\u89d2\u8272\u8bb0\u5f55\u8bc1\u636e\uff1a\n\n- \u5148\u89c2\u5bdf\uff0c\u518d\u89e3\u91ca\uff1b\u4e8b\u5b9e\u4e0e\u63a8\u65ad\u5206\u5f00\u3002\n- \u9664\u4e86\u201c\u53d1\u751f\u4e86\u4ec0\u4e48\u201d\uff0c\u8fd8\u8981\u8ba9\u4e0b\u6e38\u770b\u89c1\u77e5\u8bc6\u3001\u6743\u529b\u3001\u76ee\u6807\u3001\u5173\u7cfb\u6216\u60c5\u7eea\u5728\u54ea\u4e00\u523b\u53d8\u5316\u3002\n- \u6807\u51fa\u7531\u8c01\u7684 POV \u627f\u8f7d\u53d8\u5316\u3001\u54ea\u4e2a\u53cd\u5e94\u6216\u8868\u6f14\u4e0d\u53ef\u66ff\u4ee3\uff0c\u4ee5\u53ca\u54ea\u91cc\u5b58\u5728\u5b8c\u6574\u53f0\u8bcd/\u52a8\u4f5c\u7684\u81ea\u7136\u526a\u8f91\u8fb9\u754c\u3002\n- \u8bc1\u636e\u4e0d\u8db3\u65f6\u4fdd\u7559\u4e0d\u786e\u5b9a\u6027\uff0c\u4e0d\u5236\u9020\u620f\u5267\u7ed3\u8bba\u3002\n\n## 2. \u5904\u7406\u9636\u6bb5\n\n1. **\u573a\u666f\u68c0\u6d4b**\uff1a\u5199 `scenes.json`\uff0c\u5305\u542b\u5207\u70b9\u3001\u65f6\u957f\u548c\u5e9f\u7247\u6bb5\u8fc7\u6ee4\u7ed3\u679c\u3002\n2. **\u62bd\u5e27**\uff1a\u4e3a\u89c6\u89c9\u5206\u6790\u63d0\u53d6\u4ee3\u8868\u5e27\u3002\n3. **ASR**\uff1a\u901a\u8fc7 `mimo-v2.5-asr` \u5199\u7c97\u5206\u6bb5\u5bf9\u767d `asr_result.json`\uff0c\u5e76\u5199\n   `asr_timing_evidence.json` \u8bf4\u660e\u53ef\u7528\u6027\u3001\u6709\u9650\u65f6\u95f4\u7cbe\u5ea6\u4e0e\u6587\u672c\u4fee\u6b63\u6765\u6e90\u3002\n4. **\u9759\u97f3\u68c0\u6d4b**\uff1a\u5199 `silence_periods.json`\uff0c\u6807\u6ce8\u5b89\u9759\u7a97\u53e3\u4e0e `has_speech`\u3002\n5. **VLM \u89c2\u5bdf**\uff1a\u5199 `vlm_analysis.json`\uff0c\u5305\u542b\u573a\u666f\u63cf\u8ff0\u3001\u6df1\u5c42\u5206\u6790\u548c `frame_facts`\u3002\n6. **\u65f6\u95f4\u7ebf\u878d\u5408\u4e0e\u521b\u4f5c brief**\uff1a\u5199 `timeline_fusion.json`\u3001`asr_writing_chunks.json` \u548c `agent_narration_brief.md`\u3002\n\n\u5404\u9636\u6bb5\u53ea\u6709\u5728\u8f93\u51fa\u4ea7\u7269\u4e0e provenance sidecar \u540c\u65f6\u5339\u914d\u5f53\u524d\u89c6\u9891\u53ca\u5f71\u54cd\u7ed3\u679c\u7684\u8bbe\u7f6e\u65f6\u624d\u4f1a\u590d\u7528\uff1b`--force` \u5f3a\u5236\u91cd\u7b97\u3002\n\n## 3. \u73af\u5883\u8981\u6c42\n\n```bash\n# ffmpeg: brew install ffmpeg | apt install ffmpeg | choco install ffmpeg\nexport MIMO_API_KEY=***\n```\n\nASR \u4f7f\u7528 `mimo-v2.5-asr`\uff1bVLM \u4f7f\u7528 `mimo-v2.5`\u3002`--skip-asr` \u53ef\u8df3\u8fc7\u5bf9\u767d\u8f6c\u5199\uff0c\u4f46\u5b8c\u6574\u7406\u89e3\u4ecd\u9700\u8981 `MIMO_API_KEY` \u8fd0\u884c VLM\u3002`--mimo-video-overview` \u53ef\u5f00\u542f\u6309\u573a\u666f\u5757\u7684\u89c6\u9891\u6982\u89c8\u3002\u672a\u8bbe\u7f6e key \u65f6\u91cd\u8dd1\u4f1a\u590d\u7528\u5df2\u7f13\u5b58\u7684\u8f6c\u5199\u3001\u753b\u9762\u5206\u6790\u3001\u6982\u89c8\u4e0e\u6545\u4e8b\u7d22\u5f15\uff08key \u51b3\u5b9a\u7684\u9ed8\u8ba4 endpoint \u4e0d\u53c2\u4e0e\u6bd4\u5bf9\uff09\uff1b\u9700\u8981\u8bf7\u6c42\u6a21\u578b\u7684 consolidation \u8bb0\u4e3a `skipped_no_key`\uff0c\u4e0d\u53d1\u8bf7\u6c42\u3002\u7f13\u5b58\u5bf9\u4e0d\u4e0a\uff08\u4f8b\u5982\u590d\u5236 work_dir \u65f6\u6ca1\u4fdd\u7559\u6587\u4ef6\u65f6\u95f4\uff09\u800c\u5df2\u6709\u8f6c\u5199\u65f6\uff0c\u8fd0\u884c\u505c\u4e0b\u5e76\u4fdd\u7559\u8f6c\u5199\uff1a\u7528 `cp -p` / `cp -Rp` / `rsync -t` \u4fdd\u7559\u65f6\u95f4\u91cd\u65b0\u590d\u5236\uff0c\u6216\u8bbe\u7f6e key \u540e\u91cd\u8dd1\uff08\u4f1a\u91cd\u65b0\u8f6c\u5199\uff09\u3002\u4e0d\u8981\u7528 `--skip-asr` \u7ed5\u8fc7\uff0c\u5b83\u4f1a\u628a\u73b0\u6709\u8f6c\u5199\u66ff\u6362\u6210 `[]`\u3002\n\n\u82e5 `work_dir/background_research.json` \u5b58\u5728\uff0c\u672c\u6280\u80fd\u4f1a\u628a\u5267\u60c5\u6897\u6982\u548c\u89d2\u8272\u540d\u6298\u5165 VLM \u4e0a\u4e0b\u6587\uff1b`--context` \u53ef\u8865\u5145\u4e00\u6761\u7b80\u77ed\u63d0\u793a\u3002\n\n### \u8fdc\u7a0b\u8c03\u7528\u4e0e\u672c\u5730\u8bfb\u53d6\n\nASR \u628a\u5206\u7a97\u97f3\u9891\u53d1\u7ed9 MiMo ASR\uff1b\u9010\u573a\u666f VLM \u628a\u4ee3\u8868\u5e27\u3001\u5e27\u65f6\u95f4\u548c\u5206\u6790\u63d0\u793a\u53d1\u7ed9 MiMo\uff1b\u542f\u7528 `--mimo-video-overview` \u65f6\uff0c\u8fd8\u4f1a\u628a\u6309\u573a\u666f\u622a\u53d6\u7684\u89c6\u9891\u7247\u6bb5\u53d1\u7ed9 MiMo \u89c6\u9891\u6a21\u578b\u3002\u6545\u4e8b\u7d22\u5f15\u6c47\u603b\u548c\u53ef\u9009 ASR \u6e05\u6d17\u53d1\u9001\u7684\u662f\u753b\u9762\u63cf\u8ff0\u3001\u8f6c\u5199\u4e0e\u80cc\u666f\u8d44\u6599\u6587\u5b57\u3002\n\n\u8bf7\u6c42\u4f7f\u7528 MiMo \u7684 chat \u63a5\u53e3\u3002`MIMO_API_URL` \u53ef\u6307\u5b9a\u5730\u5740\uff1bASR \u4e0e\u89c6\u9891\u6982\u89c8\u53ef\u5206\u522b\u7528 `MIMO_ASR_API_URL`\u3001`MIMO_VIDEO_API_URL` \u8986\u76d6\uff0c\u51ed\u636e\u540c\u6837\u652f\u6301\u5404\u81ea\u7684 key\u3002\u672a\u8986\u76d6\u65f6\uff0c\u6309\u91cf key \u4f7f\u7528 `https://api.xiaomimimo.com/v1`\uff0cToken Plan key \u4f7f\u7528 `MIMO_TOKEN_PLAN_CLUSTER` \u9009\u62e9\u7684\u96c6\u7fa4\u3002\u9010\u573a\u666f\u5e27\u5206\u6790\u4e0e\u6c47\u603b\u4f7f\u7528 `MIMO_API_KEY`\u3002\n\n`--skip-asr` \u53ea\u5173\u95ed\u97f3\u9891\u8f6c\u5199\uff0c`--no-consolidate` \u53ea\u5173\u95ed\u7d22\u5f15\u6c47\u603b\uff1b\u5b83\u4eec\u4e0d\u4f1a\u5173\u95ed VLM\u3002\u9700\u8981\u53ea\u5728\u672c\u5730\u5904\u7406\u5df2\u6709\u4ea7\u7269\u65f6\uff0c\u4f7f\u7528 `--brief-only` \u6216 `--edited-storyboard-only`\uff0c\u8fd9\u4e24\u6761\u8def\u5f84\u4e0d\u8c03\u7528 API\u3002\n\n\u4e0b\u9762\u7684 `scripts/...` \u5747\u76f8\u5bf9\u4e8e\u672c\u6280\u80fd\u76ee\u5f55\u3002\u82e5\u6267\u884c\u5668\u4ece\u4ed3\u5e93\u6839\u76ee\u5f55\u542f\u52a8\uff0c\u8bf7\u7ed9\u811a\u672c\u8def\u5f84\u52a0\u4e0a\u672c\u6280\u80fd\u7684\u7edd\u5bf9\u76ee\u5f55\u3002\n\n## 4. \u8fd0\u884c\u547d\u4ee4\n\n```bash\npython3 scripts/understand.py <video> --work-dir <work_dir> [\u9009\u9879]\n```\n\n| \u9009\u9879 | \u9ed8\u8ba4 | \u4f5c\u7528 |\n|------|------|------|\n| `<video>` | \u5fc5\u586b | \u6e90\u89c6\u9891 |\n| `--work-dir` | \u5fc5\u586b | \u4ea7\u7269\u76ee\u5f55\uff1b\u4e0d\u5b58\u5728\u65f6\u521b\u5efa |\n| `--context \"...\"` | \u7a7a | \u8865\u5145\u7ed9 VLM \u7684\u7b80\u77ed\u4e0a\u4e0b\u6587\uff08\u8282\u76ee\u540d\u3001\u89d2\u8272\u540d\uff09\uff0c\u4e0e `background_research.json` \u5408\u5e76 |\n| `--scene-threshold` | `0.1` | \u573a\u666f\u68c0\u6d4b\u9608\u503c |\n| `--style` | `\u7eaa\u5f55\u7247` | \u5199\u8fdb\u521b\u4f5c\u7b80\u62a5\u7684\u89e3\u8bf4\u98ce\u683c |\n| `--edit-mode full\\|cut` | \u4e0d\u8bbe | \u5199\u8fdb\u7b80\u62a5\u7684 recap \u6a21\u5f0f\uff1b`cut` \u65f6\u6309\u526a\u540e\u65f6\u957f\u4f30\u7b97\u65c1\u767d\u9884\u7b97\uff0c\u5df2\u6709 `edited_source.mp4` \u65f6\u53e5\u672b\u951a\u70b9\u6539\u7528\u526a\u540e\u65f6\u95f4 |\n| `--target-duration` | \u4e0d\u8bbe | \u5199\u8fdb\u7b80\u62a5\u7684 cut \u76ee\u6807\u65f6\u957f\uff1b\u5c1a\u65e0 `clip_plan_validated.json` \u65f6\u7528\u5b83\u4f30\u7b97\u65c1\u767d\u9884\u7b97 |\n| `--skip-asr` | \u5173 | \u4e0d\u8f6c\u5199\u5bf9\u767d\uff0c\u628a `asr_result.json` \u5199\u6210 `[]`\uff08\u5df2\u6709\u8f6c\u5199\u4f1a\u88ab\u8986\u76d6\uff09\uff0cASR \u8bc1\u636e\u6807\u4e3a\u663e\u5f0f\u8df3\u8fc7 |\n| `--mimo-video-overview` | \u5173 | \u6309\u573a\u666f\u5757\u8fd0\u884c MiMo \u89c6\u9891\u6982\u89c8\uff0c\u5e76\u4f5c\u4e3a\u9010\u573a\u666f\u4e3b\u63cf\u8ff0 |\n| `--force` | \u5173 | \u5ffd\u7565\u7f13\u5b58\uff0c\u5168\u90e8\u91cd\u7b97 |\n| `--brief-only` | \u5173 | \u53ea\u7528\u73b0\u6709\u4ea7\u7269\u91cd\u5efa `agent_narration_brief.md`\uff0c\u4e0d\u62bd\u5e27\u3001\u4e0d\u8c03 API |\n| `--edited-storyboard-only` | \u5173 | \u53ea\u6309 `clip_plan_validated.json` \u5199\u526a\u540e\u65f6\u95f4\u7ebf `storyboard/edited_storyboard.*`\uff0c\u5e76\u5728\u5df2\u6709\u7684 `agent_narration_brief.md` \u9876\u90e8\u52a0 storyboard \u6307\u5f15\uff1b\u591a\u6e90\u8ba1\u5212\uff08\u5e26 `sources`\uff09\u4ece\u5404\u6765\u6e90 `source_work_dir` \u7684 `frames/` \u6309\u5176 `frames_manifest.json` \u7684 fps \u53d6\u5e27\uff0ctile \u6807 `S1`/`S2`\u2026\uff1b\u4e0d\u62bd\u5e27\u3001\u4e0d\u8c03 API\uff0c\u6545\u4e8b\u677f\u751f\u6210\u5931\u8d25\u53ea\u8bb0\u65e5\u5fd7\u3002\u4e0e `--brief-only` \u4e92\u65a5 |\n| `--consolidate` / `--no-consolidate` | \u5f00 | \u751f\u6210\u5168\u5c40\u6545\u4e8b\u7d22\u5f15 `understanding_index.*` |\n| `--consolidate-asr` | \u5173 | \u53e6\u5916\u6e05\u6d17 ASR \u6587\u672c\uff0c\u5199 `asr_clean.json` |\n\n## 5. \u8f93\u51fa\u5951\u7ea6\n\n\u9ed8\u8ba4\u8fd0\u884c\u5199\u51fa\u4e0b\u8868\u4ea7\u7269\uff1b\u5404\u9636\u6bb5\u4ea7\u7269\u65c1\u7684 `*.meta.json` \u662f\u7f13\u5b58 provenance sidecar\u3002\n\n| \u6587\u4ef6 | \u5185\u5bb9 |\n|------|------|\n| `frames/frame_*.jpg`\u3001`frames_manifest.json` | \u6309 fps \u62bd\u51fa\u7684\u5e27\u53ca\u5176\u6e05\u5355 |\n| `scenes.json` | \u573a\u666f\u5207\u70b9\u3001\u8d77\u6b62\u65f6\u95f4\u4e0e\u65f6\u957f |\n| `audio.wav` | 16 kHz \u5355\u58f0\u9053\u97f3\u9891\uff0c\u4f9b ASR\u3001\u9759\u97f3\u68c0\u6d4b\u4e0e\u53e5\u672b\u951a\u70b9\u4f7f\u7528 |\n| `asr_result.json` | `[{start, end, text}]` \u65f6\u95f4\u6233\u5bf9\u767d |\n| `asr_timing_evidence.json` | ASR \u53ef\u7528\u6027\u72b6\u6001\u3001\u7c97\u7a97\u53e3\u7cbe\u5ea6\u3001glossary \u524d\u540e\u6587\u672c\uff0c\u4ee5\u53ca\u5b83\u6240\u63cf\u8ff0\u7684\u6e90\u89c6\u9891/\u97f3\u9891/\u7ed3\u679c\u6587\u4ef6\uff08\u8def\u5f84\u5b58\u5728\u6027 + size/mtime\uff09 |\n| `silence_periods.json` | `[{start, end, duration, has_speech}]` \u5b89\u9759\u7a97\u53e3 |\n| `speech_boundary_anchors.json` | ASR \u53e5\u672b\u6807\u70b9\u5bf9\u9f50\u5230\u77ed\u505c\u987f\u7684\u53e5\u672b\u951a\u70b9\uff1b\u7f3a\u97f3\u9891\u6216 ASR \u65f6 `status: unavailable` |\n| `vlm_analysis.json` | \u9010\u573a\u666f\u63cf\u8ff0\u3001\u6df1\u5c42\u5206\u6790\u4e0e `frame_facts` |\n| `mimo_video_overview.status.json` | \u89c6\u9891\u6982\u89c8\u72b6\u6001\uff08\u672a\u542f\u7528\u65f6\u4e3a `disabled`\uff09\uff1b\u542f\u7528\u6210\u529f\u53e6\u5199 `mimo_video_overview.json` |\n| `understanding_index.json`\u3001`understanding_index.md` | \u5168\u5c40\u6545\u4e8b\u7d22\u5f15\uff08`--no-consolidate` \u65f6\u4e0d\u5199\uff09 |\n| `consolidation.status.json` | \u6545\u4e8b\u7d22\u5f15\u4e0e ASR \u6e05\u6d17\u7684\u8fd0\u884c\u72b6\u6001 |\n| `storyboard/source_storyboard.{json,jpg}` | \u539f\u7247\u65f6\u95f4\u7ebf\u8054\u7cfb\u8868\uff08\u8d85\u9875\u65f6\u7eed\u5199 `_001.jpg` \u7b49\uff09\uff1b\u5df2\u6709 `clip_plan_validated.json` \u65f6\u53e6\u5199 `edited_storyboard.*`\uff1b`STORYBOARD=0` \u5173\u95ed |\n| `timeline_fusion.json` | VLM\u3001ASR \u4e0e\u9759\u97f3\u4fe1\u606f\u7684\u7edf\u4e00\u65f6\u95f4\u7ebf |\n| `asr_writing_chunks.json` | \u6309\u53e5\u754c\u548c\u573a\u666f\u5207\u5206\u7684 ASR \u5199\u4f5c\u5757 |\n| `agent_narration_brief.md` | Agent \u9996\u5148\u9605\u8bfb\u7684\u521b\u4f5c\u7b80\u62a5 |\n\n\u540e\u7eed\u5199\u4f5c\u9636\u6bb5\u6839\u636e\u521b\u4f5c\u7b80\u62a5\u4e0e\u7d22\u5f15\u5236\u5b9a\u65b9\u6848\u5e76\u5199 `narration.json`\u3002\n\n## 6. \u53c2\u8003\u8d44\u6599\n\n- \u80cc\u666f\u8c03\u7814\uff1a`references/research-guide.md`\uff0c\u4ea7\u51fa `background_research.json`\u3002\n- JSON \u7ed3\u6784\uff1a`references/data-schema.md`\u3002\n\n## 7. \u80fd\u529b\u8fb9\u754c\n\n- \u4e0d\u5199\u89e3\u8bf4\u8bcd\uff0c\u4e5f\u4e0d\u505a\u89e3\u8bf4\u8bc4\u5206\uff1b\u53ea\u8d1f\u8d23\u751f\u6210\u7406\u89e3\u7d22\u5f15\u4e0e\u521b\u4f5c\u7b80\u62a5\u3002\n- \u4e0d\u7f16\u9020\u4fe1\u53f7\u65e0\u6cd5\u652f\u6301\u7684\u5267\u60c5\uff1b\u5f53 ASR / VLM \u8fc7\u8584\u65f6\u8f93\u51fa\u7d20\u6750\u8b66\u544a\u3002\n- MiMo ASR \u7684 `start/end` \u662f\u56fa\u5b9a\u5206\u7247\u5f62\u6210\u7684**\u7c97\u7a97\u53e3**\uff0c\u4e0d\u662f\u8bcd\u7ea7\u5bf9\u9f50\uff1b\u7a7a\u6587\u672c\u53ea\u8868\u793a\u539f\u56e0\u672a\u77e5\uff0c\n  \u4e0d\u80fd\u5f53\u4f5c\u5df2\u8bc1\u5b9e\u9759\u97f3\u3002`asr_timing_evidence.json` \u7684\u72b6\u6001\u5b57\u6bb5\u89c1 `references/data-schema.md`\u3002\n","usage":"\u8bfb\u53d6 SKILL.md\uff0c\u5e76\u6309\u9700\u8bfb\u53d6\u914d\u5957\u6587\u4ef6\uff1b\u4e0b\u8f7d ZIP \u83b7\u53d6\u5b8c\u6574\u5305\u3002\u6280\u80fd\u662f\u4efb\u52a1\u65b9\u6cd5\uff0c\u4e0d\u6388\u4e88\u65b0\u589e\u8d26\u6237\u3001\u4ed8\u8d39\u6216\u7cfb\u7edf\u6743\u9650\u3002"}
