[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6D0AX3ZWYKEEH79EEW4K6WC":3,"skill-same:01M6D0AX3ZWYKEEH79EEW4K6WC":79},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":28,"publishedAt":46,"linkOnly":47,"markdownContent":48,"markdownBody":49,"frontMatter":50,"refName":58,"commitSha":59,"skillDirectory":60,"forksCount":61,"repositoryHtmlUrl":62,"repositoryHomepageUrl":63,"ownerHtmlUrl":64,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":65,"install":66,"resources":78},"01M6D0AX3ZWYKEEH79EEW4K6WC","claude_code_delegation",null,"Delegate coding, review, and refactor tasks to Claude Code CLI through Row-Bot's approval-gated shell workflow.","siddsachar\u002Frow-bot","siddsachar","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F50956096?v=4","01M6D0AG8BTX8VNFY5N3F2Q36T","Siddharth Sachar","01M6D0AGD3C7BZ6NAW9PWVKM3A",1687,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Fbundled_skills\u002Fclaude_code_delegation\u002FSKILL.md",[18,23],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":26,"taxonomyTypeDesc":27},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45],"claude code delegation","code","delegation","refactor","true","claude code cli","gated","approval","delegate","coding","shell","bot","tasks","automation","workflow","cli","review","2026-11-10 10:12:30",false,"---\nname: claude_code_delegation\ndisplay_name: Claude Code Delegation\nicon: \"?\"\ndescription: Delegate coding, review, and refactor tasks to Claude Code CLI through Row-Bot's approval-gated shell workflow.\nenabled_by_default: true\nversion: \"1.0\"\ntags:\n  - coding\n  - claude\n  - delegation\n  - review\n  - automation\nauthor: Row-Bot\n---\n\nWhen the user asks to **use Claude Code**, **coordinate Claude Code**, **delegate coding work**, **have another coding agent implement\u002Freview\u002Frefactor something**, or compares Row-Bot with Hermes-style Claude Code orchestration, use this workflow.\n\nYou are the coordinator. Claude Code is the delegated coding worker. Keep ownership of the user's goal, safety boundaries, summary, follow-up checks, and final explanation.\n\n## Core Principles\n\n1. **Prefer print mode first** — Use `claude -p` for most tasks. It runs one bounded task, returns output, and exits. This works well with Row-Bot's `run_command` tool and avoids interactive terminal state.\n2. **Keep Row-Bot in charge** — Do not hand off the whole conversation. Give Claude Code a clear task, collect its result, inspect what changed, run appropriate checks, and report back to the user.\n3. **Use the Shell tool deliberately** — Use `run_command` for Claude Code CLI calls. Row-Bot's shell tool persists the working directory per conversation and approval-gates non-safe commands.\n4. **Set the project directory explicitly** — Before delegating, confirm or set the repo directory with `pwd` \u002F `Get-Location` and `cd` \u002F `Set-Location`. Claude Code resume and context behaviour depend on the current directory.\n5. **Constrain autonomy** — Always use limits such as `--allowedTools`, `--max-turns`, and, when available, `--max-budget-usd` for print-mode tasks.\n6. **Avoid permission bypass** — Do not use `--dangerously-skip-permissions` unless the user explicitly asks and understands that Claude Code may edit files and run commands without per-action prompts.\n7. **Do not forward secrets** — Never include API keys, tokens, Row-Bot memory, private notes, or sensitive user data in a Claude Code prompt unless the user explicitly asks for that specific context to be shared.\n\n## Prerequisite Checks\n\nBefore the first delegation in a project or session, check whether Claude Code is available and authenticated:\n\n```text\nclaude --version\nclaude auth status\nclaude doctor\n```\n\nIf these fail, explain the setup path briefly:\n\n```text\nnpm install -g @anthropic-ai\u002Fclaude-code\nclaude auth login\n```\n\nOn Windows, native print-mode CLI usage may work if `claude` is installed on PATH. For interactive\u002Ftmux-style orchestration, recommend WSL2 unless native terminal support has been verified.\n\n## Delegation Workflow\n\n1. **Understand the task** — Restate the requested coding outcome, target repo\u002Fpath, and expected verification. Ask one clarifying question only if the target or risk level is unclear.\n2. **Inspect local state first** — Use safe shell commands such as `git status`, `git diff`, and directory listing commands before delegation. Notice uncommitted user changes and do not overwrite them.\n3. **Choose capability scope** — Pick the narrowest Claude Code tool set for the task:\n   - Review\u002Fexplain only: `--allowedTools \"Read\"`\n   - Edit existing files: `--allowedTools \"Read,Edit\"`\n   - Create files\u002Ftests: `--allowedTools \"Read,Edit,Write\"`\n   - Run tests\u002Fbuilds: `--allowedTools \"Read,Edit,Write,Bash\"`\n4. **Ask for approval when needed** — A Claude Code run that can edit files or run shell commands is write-capable delegation. If the user did not already authorize that, confirm before running it.\n5. **Run Claude Code in print mode** — Use a bounded command with a concise prompt, explicit limits, and JSON output when useful.\n6. **Inspect the result** — After Claude Code exits, run `git status` and inspect the diff. Do not assume the delegated agent did the right thing.\n7. **Verify** — Run targeted tests, lint, type checks, or smoke checks appropriate to the change. If verification is expensive or unclear, ask before running broad suites.\n8. **Report clearly** — Summarize what Claude Code changed, what you verified, any failures or uncertainties, and recommended next steps.\n\n## Print-Mode Command Patterns\n\nUse these as patterns, adapting quoting for the user's shell. In Row-Bot, it is often cleaner to set the directory first, then run `claude` as a separate command so the shell session cwd persists.\n\n### Status and Setup\n\n```text\nclaude --version\nclaude auth status\nclaude doctor\n```\n\n### Read-Only Review\n\n```text\nclaude -p \"Review the current repository changes for bugs, security issues, and missing tests. Do not edit files. Return findings with file paths and verification suggestions.\" --allowedTools \"Read\" --max-turns 3 --output-format json\n```\n\n### Focused Edit\n\n```text\nclaude -p \"Implement the requested fix in the smallest safe change. Preserve existing style. After editing, summarize changed files and tests to run.\" --allowedTools \"Read,Edit\" --max-turns 8 --max-budget-usd 2 --output-format json\n```\n\n### Feature Work With Tests\n\n```text\nclaude -p \"Implement this feature and add focused tests. Prefer minimal, idiomatic changes. Run the relevant tests if safe. Return a summary, files changed, tests run, and any remaining risks.\" --allowedTools \"Read,Edit,Write,Bash\" --max-turns 12 --max-budget-usd 4 --output-format json\n```\n\n### Diff Review\n\n```text\ngit diff --stat\ngit diff | claude -p \"Review this diff for correctness, regressions, security issues, and missing tests. Do not edit files.\" --allowedTools \"Read\" --max-turns 1\n```\n\nIf the diff is large, save a focused patch file for Claude Code to inspect, but avoid including secrets or unrelated local changes. Remember that Claude Code cannot see Row-Bot's previous tool output unless you pass the relevant context into the Claude Code prompt, stdin, files, or repository state.\n\n## Prompt Construction\n\nGive Claude Code a compact brief:\n\n```text\nGoal: \u003Cspecific outcome>\nRepo\u002Fpath: \u003Ccurrent directory or subpath>\nConstraints: preserve user changes, follow existing patterns, avoid broad refactors\nAllowed work: \u003Cread-only | edit existing files | create tests | run tests>\nVerification: \u003Cspecific commands or ask it to identify tests>\nOutput: summary, files changed, tests run, remaining risks\n```\n\nDo not ask Claude Code to make commits, push branches, publish releases, delete data, rotate secrets, or run production-affecting commands unless the user explicitly requested that operation.\n\n## Session Continuation\n\nWhen using `--output-format json`, capture the returned `session_id` if Claude Code provides one. Use resume only when continuing the same task in the same project directory:\n\n```text\nclaude -p \"Continue from the previous session and address the remaining test failure.\" --resume \u003Csession_id> --max-turns 5 --allowedTools \"Read,Edit,Bash\"\n```\n\nUse `--continue` only when you are sure the most recent Claude Code session in the current directory is the one you want. If unsure, start a fresh print-mode task instead.\n\n## Worktree Mode\n\nFor larger or risky edits, prefer isolating Claude Code's work in a branch or worktree when available:\n\n```text\nclaude -p \"Implement the feature in an isolated worktree and summarize the diff.\" --worktree row-bot-delegation-task --allowedTools \"Read,Edit,Write,Bash\" --max-turns 12 --max-budget-usd 4\n```\n\nAfterward, inspect the worktree diff before merging or copying changes back. Do not merge or delete worktrees without user approval.\n\n## Interactive Mode Is Advanced\n\nInteractive Claude Code is useful for long, multi-turn coding sessions, Claude Code slash commands, and human-in-the-loop exploration. It is also more fragile.\n\nUse interactive mode only when print mode is insufficient. Prefer macOS\u002FLinux\u002FWSL2 with `tmux` for reliable monitoring:\n\n```text\ntmux new-session -d -s claude-work -x 140 -y 40\ntmux send-keys -t claude-work 'cd \u002Fpath\u002Fto\u002Fproject; claude' Enter\ntmux capture-pane -t claude-work -p -S -50\ntmux send-keys -t claude-work 'Now add focused regression tests' Enter\ntmux send-keys -t claude-work '\u002Fexit' Enter\ntmux kill-session -t claude-work\n```\n\nRules for interactive mode:\n\n- Prefer WSL2 on Windows for tmux-based orchestration.\n- Monitor with `tmux capture-pane` before assuming the session is stuck.\n- Look for a prompt or clear waiting state before sending follow-up input.\n- Clean up tmux sessions when done.\n- Do not leave background coding agents running silently.\n\n## Safety Boundaries\n\n- Treat Claude Code as a powerful external actor that can edit the repo and spend API quota.\n- Use `--allowedTools` every time.\n- Use `--max-turns` for print mode every time.\n- Use `--max-budget-usd` for non-trivial tasks when available.\n- Avoid broad filesystem access with `--add-dir` unless required.\n- Avoid `--permission-mode bypassPermissions` and `--dangerously-skip-permissions` by default.\n- Do not run destructive git commands, force pushes, deploys, production migrations, or secret-handling tasks through Claude Code without explicit user approval.\n- Respect Row-Bot background workflow safety modes: write-capable Claude Code runs should not happen in unattended background tasks unless the user selected an approval level that allows it for that task (Ask, reviewed Auto, or Hands-off).\n\n## After Delegation\n\nAlways follow up with Row-Bot-native review:\n\n1. Run `git status`.\n2. Inspect changed files or the diff.\n3. Run relevant tests\u002Fchecks if appropriate.\n4. Tell the user:\n   - what Claude Code attempted,\n   - what changed,\n   - what passed or failed,\n   - what still needs human review.\n\nIf Claude Code fails, times out, hits a budget\u002Fturn limit, or produces ambiguous output, do not keep retrying blindly. Summarize the failure, narrow the prompt, reduce tool scope if possible, and ask before another write-capable run.","When the user asks to **use Claude Code**, **coordinate Claude Code**, **delegate coding work**, **have another coding agent implement\u002Freview\u002Frefactor something**, or compares Row-Bot with Hermes-style Claude Code orchestration, use this workflow.\n\nYou are the coordinator. Claude Code is the delegated coding worker. Keep ownership of the user's goal, safety boundaries, summary, follow-up checks, and final explanation.\n\n## Core Principles\n\n1. **Prefer print mode first** — Use `claude -p` for most tasks. It runs one bounded task, returns output, and exits. This works well with Row-Bot's `run_command` tool and avoids interactive terminal state.\n2. **Keep Row-Bot in charge** — Do not hand off the whole conversation. Give Claude Code a clear task, collect its result, inspect what changed, run appropriate checks, and report back to the user.\n3. **Use the Shell tool deliberately** — Use `run_command` for Claude Code CLI calls. Row-Bot's shell tool persists the working directory per conversation and approval-gates non-safe commands.\n4. **Set the project directory explicitly** — Before delegating, confirm or set the repo directory with `pwd` \u002F `Get-Location` and `cd` \u002F `Set-Location`. Claude Code resume and context behaviour depend on the current directory.\n5. **Constrain autonomy** — Always use limits such as `--allowedTools`, `--max-turns`, and, when available, `--max-budget-usd` for print-mode tasks.\n6. **Avoid permission bypass** — Do not use `--dangerously-skip-permissions` unless the user explicitly asks and understands that Claude Code may edit files and run commands without per-action prompts.\n7. **Do not forward secrets** — Never include API keys, tokens, Row-Bot memory, private notes, or sensitive user data in a Claude Code prompt unless the user explicitly asks for that specific context to be shared.\n\n## Prerequisite Checks\n\nBefore the first delegation in a project or session, check whether Claude Code is available and authenticated:\n\n```text\nclaude --version\nclaude auth status\nclaude doctor\n```\n\nIf these fail, explain the setup path briefly:\n\n```text\nnpm install -g @anthropic-ai\u002Fclaude-code\nclaude auth login\n```\n\nOn Windows, native print-mode CLI usage may work if `claude` is installed on PATH. For interactive\u002Ftmux-style orchestration, recommend WSL2 unless native terminal support has been verified.\n\n## Delegation Workflow\n\n1. **Understand the task** — Restate the requested coding outcome, target repo\u002Fpath, and expected verification. Ask one clarifying question only if the target or risk level is unclear.\n2. **Inspect local state first** — Use safe shell commands such as `git status`, `git diff`, and directory listing commands before delegation. Notice uncommitted user changes and do not overwrite them.\n3. **Choose capability scope** — Pick the narrowest Claude Code tool set for the task:\n   - Review\u002Fexplain only: `--allowedTools \"Read\"`\n   - Edit existing files: `--allowedTools \"Read,Edit\"`\n   - Create files\u002Ftests: `--allowedTools \"Read,Edit,Write\"`\n   - Run tests\u002Fbuilds: `--allowedTools \"Read,Edit,Write,Bash\"`\n4. **Ask for approval when needed** — A Claude Code run that can edit files or run shell commands is write-capable delegation. If the user did not already authorize that, confirm before running it.\n5. **Run Claude Code in print mode** — Use a bounded command with a concise prompt, explicit limits, and JSON output when useful.\n6. **Inspect the result** — After Claude Code exits, run `git status` and inspect the diff. Do not assume the delegated agent did the right thing.\n7. **Verify** — Run targeted tests, lint, type checks, or smoke checks appropriate to the change. If verification is expensive or unclear, ask before running broad suites.\n8. **Report clearly** — Summarize what Claude Code changed, what you verified, any failures or uncertainties, and recommended next steps.\n\n## Print-Mode Command Patterns\n\nUse these as patterns, adapting quoting for the user's shell. In Row-Bot, it is often cleaner to set the directory first, then run `claude` as a separate command so the shell session cwd persists.\n\n### Status and Setup\n\n```text\nclaude --version\nclaude auth status\nclaude doctor\n```\n\n### Read-Only Review\n\n```text\nclaude -p \"Review the current repository changes for bugs, security issues, and missing tests. Do not edit files. Return findings with file paths and verification suggestions.\" --allowedTools \"Read\" --max-turns 3 --output-format json\n```\n\n### Focused Edit\n\n```text\nclaude -p \"Implement the requested fix in the smallest safe change. Preserve existing style. After editing, summarize changed files and tests to run.\" --allowedTools \"Read,Edit\" --max-turns 8 --max-budget-usd 2 --output-format json\n```\n\n### Feature Work With Tests\n\n```text\nclaude -p \"Implement this feature and add focused tests. Prefer minimal, idiomatic changes. Run the relevant tests if safe. Return a summary, files changed, tests run, and any remaining risks.\" --allowedTools \"Read,Edit,Write,Bash\" --max-turns 12 --max-budget-usd 4 --output-format json\n```\n\n### Diff Review\n\n```text\ngit diff --stat\ngit diff | claude -p \"Review this diff for correctness, regressions, security issues, and missing tests. Do not edit files.\" --allowedTools \"Read\" --max-turns 1\n```\n\nIf the diff is large, save a focused patch file for Claude Code to inspect, but avoid including secrets or unrelated local changes. Remember that Claude Code cannot see Row-Bot's previous tool output unless you pass the relevant context into the Claude Code prompt, stdin, files, or repository state.\n\n## Prompt Construction\n\nGive Claude Code a compact brief:\n\n```text\nGoal: \u003Cspecific outcome>\nRepo\u002Fpath: \u003Ccurrent directory or subpath>\nConstraints: preserve user changes, follow existing patterns, avoid broad refactors\nAllowed work: \u003Cread-only | edit existing files | create tests | run tests>\nVerification: \u003Cspecific commands or ask it to identify tests>\nOutput: summary, files changed, tests run, remaining risks\n```\n\nDo not ask Claude Code to make commits, push branches, publish releases, delete data, rotate secrets, or run production-affecting commands unless the user explicitly requested that operation.\n\n## Session Continuation\n\nWhen using `--output-format json`, capture the returned `session_id` if Claude Code provides one. Use resume only when continuing the same task in the same project directory:\n\n```text\nclaude -p \"Continue from the previous session and address the remaining test failure.\" --resume \u003Csession_id> --max-turns 5 --allowedTools \"Read,Edit,Bash\"\n```\n\nUse `--continue` only when you are sure the most recent Claude Code session in the current directory is the one you want. If unsure, start a fresh print-mode task instead.\n\n## Worktree Mode\n\nFor larger or risky edits, prefer isolating Claude Code's work in a branch or worktree when available:\n\n```text\nclaude -p \"Implement the feature in an isolated worktree and summarize the diff.\" --worktree row-bot-delegation-task --allowedTools \"Read,Edit,Write,Bash\" --max-turns 12 --max-budget-usd 4\n```\n\nAfterward, inspect the worktree diff before merging or copying changes back. Do not merge or delete worktrees without user approval.\n\n## Interactive Mode Is Advanced\n\nInteractive Claude Code is useful for long, multi-turn coding sessions, Claude Code slash commands, and human-in-the-loop exploration. It is also more fragile.\n\nUse interactive mode only when print mode is insufficient. Prefer macOS\u002FLinux\u002FWSL2 with `tmux` for reliable monitoring:\n\n```text\ntmux new-session -d -s claude-work -x 140 -y 40\ntmux send-keys -t claude-work 'cd \u002Fpath\u002Fto\u002Fproject; claude' Enter\ntmux capture-pane -t claude-work -p -S -50\ntmux send-keys -t claude-work 'Now add focused regression tests' Enter\ntmux send-keys -t claude-work '\u002Fexit' Enter\ntmux kill-session -t claude-work\n```\n\nRules for interactive mode:\n\n- Prefer WSL2 on Windows for tmux-based orchestration.\n- Monitor with `tmux capture-pane` before assuming the session is stuck.\n- Look for a prompt or clear waiting state before sending follow-up input.\n- Clean up tmux sessions when done.\n- Do not leave background coding agents running silently.\n\n## Safety Boundaries\n\n- Treat Claude Code as a powerful external actor that can edit the repo and spend API quota.\n- Use `--allowedTools` every time.\n- Use `--max-turns` for print mode every time.\n- Use `--max-budget-usd` for non-trivial tasks when available.\n- Avoid broad filesystem access with `--add-dir` unless required.\n- Avoid `--permission-mode bypassPermissions` and `--dangerously-skip-permissions` by default.\n- Do not run destructive git commands, force pushes, deploys, production migrations, or secret-handling tasks through Claude Code without explicit user approval.\n- Respect Row-Bot background workflow safety modes: write-capable Claude Code runs should not happen in unattended background tasks unless the user selected an approval level that allows it for that task (Ask, reviewed Auto, or Hands-off).\n\n## After Delegation\n\nAlways follow up with Row-Bot-native review:\n\n1. Run `git status`.\n2. Inspect changed files or the diff.\n3. Run relevant tests\u002Fchecks if appropriate.\n4. Tell the user:\n   - what Claude Code attempted,\n   - what changed,\n   - what passed or failed,\n   - what still needs human review.\n\nIf Claude Code fails, times out, hits a budget\u002Fturn limit, or produces ambiguous output, do not keep retrying blindly. Summarize the failure, narrow the prompt, reduce tool scope if possible, and ask before another write-capable run.",{"icon":51,"name":5,"tags":52,"author":54,"version":55,"description":7,"display_name":56,"enabled_by_default":57},"?",[38,53,31,45,42],"claude","Row-Bot","1.0","Claude Code Delegation",true,"main","aea3219498f535052aba82436fc9612f548a086e","bundled_skills\u002Fclaude_code_delegation",204,"https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot","https:\u002F\u002Frow-bot.ai","https:\u002F\u002Fgithub.com\u002Fsiddsachar","2026-11-15 02:36:29",{"source":8,"skillName":5,"commands":67,"prompt":77,"zipAvailable":57},[68,71,74],{"tool":69,"command":70},"npx","npx skills add siddsachar\u002Frow-bot --skill claude_code_delegation",{"tool":72,"command":73},"bunx","bunx skills add siddsachar\u002Frow-bot --skill claude_code_delegation",{"tool":75,"command":76},"pnpm","pnpm dlx skills add siddsachar\u002Frow-bot --skill claude_code_delegation","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Fbundled_skills\u002Fclaude_code_delegation\u002FSKILL.md",[],[80,101,121,132,142,157],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":46},"01M6D0EZEJNX7R2J8EXN3386BW","x_guide","Essential guidance for using X (Twitter) tools correctly — search syntax, time filtering, and engagement patterns.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fx_guide\u002FSKILL.md",[],[87,88,89,40,90,91,92,93,94,95,96,97,98,99,100],"x guide","guide","for","search","Tools","syntax","twitter","filtering","engagement","guidance","patterns","social","time","essential",{"skillNo":102,"name":103,"nameZh":6,"description":104,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":105,"categories":106,"keywords":108,"publishedAt":46},"01M6D0EWESTPTF4B1JZFVFQ49Y","workflow_guide","How workflows run (scheduled chats), how to write and change them, hand work between them, and check what a run did.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fworkflow_guide\u002FSKILL.md",[107],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[109,88,43,110,111,112,113,114,115,116,117,118,119,120],"workflow guide","write","task","how to","hand","scheduled","check","workflows","change","run","how","work",{"skillNo":122,"name":123,"nameZh":6,"description":124,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":125,"categories":126,"keywords":127,"publishedAt":46},"01M6D0ESS92MEV9V5EAHJ3N6VR","wiki_guide","Guidance for the Obsidian-compatible wiki vault.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fwiki_guide\u002FSKILL.md",[],[128,88,129,130,131,96,89],"wiki guide","wiki","obsidian","vault",{"skillNo":133,"name":134,"nameZh":6,"description":135,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":136,"categories":137,"keywords":138,"publishedAt":46},"01M6D0EQDFABHZ27HMDKYSX6V4","weather_guide","Guidance for using weather tools effectively.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fweather_guide\u002FSKILL.md",[],[139,88,140,91,141,96,89],"weather guide","weather","weather tools",{"skillNo":143,"name":144,"nameZh":6,"description":145,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":146,"categories":147,"keywords":148,"publishedAt":46},"01M6D0EMVZ8MZHAVSTMMTMBWN0","vision_guide","Guidance for using webcam, screen capture, and image analysis.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fvision_guide\u002FSKILL.md",[],[149,150,88,96,151,152,153,89,154,155,156],"vision guide","vision","analysis","screen","image analysis","capture","screen capture","image",{"skillNo":158,"name":159,"nameZh":6,"description":160,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":161,"categories":162,"keywords":166,"publishedAt":46},"01M6D0EJF8QMGWNNKD38BK24G8","video_guide","Guidance for generating videos and animating images.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fvideo_guide\u002FSKILL.md",[163],{"categoryNo":164,"categoryName":165,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000016","图像与多媒体",[88,167,168,96,169,89,170],"Video","video guide","images","videos"]