[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7DE161PD3HRPCRTNNW6V43P":3,"skill-same:01M7DE161PD3HRPCRTNNW6V43P":49},{"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":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":6,"ownerHtmlUrl":31,"ownerBlogUrl":32,"ownerDescription":33,"updatedAt":34,"install":35,"resources":48},"01M7DE161PD3HRPCRTNNW6V43P","make-bot-ui",null,"Use when building a custom UI that starts agent tasks through a webhook or local runner.","pascalandy\u002Fskills","pascalandy","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F6694151?v=4","01M7DDRR1T4AMR6J3FV6RHHV5M","Pascal Andy","01M7DDRR2B04BJ8S7QZYZV1WFK",0,"MIT","https:\u002F\u002Fgithub.com\u002Fpascalandy\u002Fskills\u002Fblob\u002Fmain\u002Fauthoring\u002Fpoteto\u002Fmake-bot-ui\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: \"make-bot-ui\"\ndescription: \"Use when building a custom UI that starts agent tasks through a webhook or local runner.\"\nkind: \"dev\"\nrole: \"helper\"\n---\n\n# Make a bot UI\n\nBuild a page whose buttons submit bounded jobs to a server. The server authenticates the request and starts an agent task. Keep credentials on the server. Read [agent runtime](..\u002Fpoteto-mode\u002Freferences\u002Fagent-runtime.md) before selecting the runner.\n\n## Select the runner\n\nDiscover the installed agent's supported automation interface. Inspect its CLI help or server API before writing the adapter. A chat interface alone is not a webhook receiver. Load `$headless` for each CLI's non-interactive flags. Its `headless.py` launcher runs blocking reviews and streams no job events.\n\n| Agent | Available integration to check |\n| --- | --- |\n| Codex | A server-owned `codex exec` process with JSON event output or a configured SDK integration |\n| Claude Code | A server-owned `claude -p` process with structured output or Agent SDK integration |\n| Pi | A supervised `pi --mode rpc` process, SDK session, or separate print-mode invocation |\n| OpenCode | Its configured server\u002FSDK session API, including asynchronous prompts where supported |\n\nUse an existing authenticated webhook service if the user already has one. Read its actual URL, authentication scheme, request format, and success semantics. Otherwise implement the local server as the adapter to the selected runner. Record whether jobs start fresh sessions or resume a specific task session. Serialize requests to a shared session or checkout.\n\nIf no callable runner is available, prepare the UI and request contract, then report that starting agent jobs is blocked. Do not claim that a button can wake the current chat or survive a restart without a configured service that does so.\n\n## Define the job contract\n\nName the allowed actions and their JSON fields. Validate input at the server boundary. Pass data separately from fixed task instructions and use process argument arrays or structured SDK calls, never a shell command assembled from request text.\n\nReturn a job ID when the server accepts work. Report queued, running, completed, or failed from the runner's actual events and exit status. An accepted HTTP request does not prove that the agent started or completed the task. Use an idempotency key if requests can be retried. Persist pending jobs only when recovery is required, and mark interrupted jobs explicitly after restart.\n\n## Configure credentials\n\nUse the environment's secret-entry facility when available. Otherwise ask the user to enter credentials directly into a server-side secret store or restricted local configuration file. Never request secret values in chat. Follow the repository's managed-file rules for configuration.\n\nKeep provider credentials and upstream webhook keys out of browser code, responses, URLs, screenshots, and logs. Use the upstream service's documented authentication headers. Do not copy authentication headers from another provider.\n\n## Serve the UI\n\nServe the page and job endpoints from one origin. Keep the agent API behind this server. Authenticate job submission, enforce the allowed action list, and restrict browser origins. Set a short timeout for accepting a request and a separate job runtime limit. A slow agent must not hold a browser request open indefinitely.\n\nPrefer a localhost listener behind an existing Tailscale proxy. If using a direct tailnet listener, bind to the node's Tailscale address and confirm the interface. Do not expose an unauthenticated agent API on all network interfaces.\n\nUse an existing online Tailscale node. Inspect `tailscale status` and `tailscale ip -4`, then report the actual reachable URL. Follow the installed version's setup workflow if the node is not connected. Authentication belongs in the user's browser or secret-entry UI, not in chat. Do not install or reconfigure another node without need.\n\n## Prove the complete path\n\nUse a harmless allowed action. Click the real page, observe the accepted job ID, then inspect the runner's start event, completion event, and expected result. Verify that failed starts and malformed input produce useful errors. If recovery or retries are supported, prove that resubmission does not execute the same job twice.\n\nReport the UI URL, selected runner, where job status is visible, and what must remain running. A local server stops providing service when its host stops unless another configured host takes over.\n\nIntegration references: [Codex non-interactive mode](https:\u002F\u002Fdevelopers.openai.com\u002Fcodex\u002Fnoninteractive), [Claude Code programmatic use](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fheadless), [Pi RPC](https:\u002F\u002Fgithub.com\u002Fbadlogic\u002Fpi-mono\u002Fblob\u002Fmain\u002Fpackages\u002Fcoding-agent\u002Fdocs\u002Frpc.md), [OpenCode server](https:\u002F\u002Fopencode.ai\u002Fdocs\u002Fserver\u002F).\n","# Make a bot UI\n\nBuild a page whose buttons submit bounded jobs to a server. The server authenticates the request and starts an agent task. Keep credentials on the server. Read [agent runtime](..\u002Fpoteto-mode\u002Freferences\u002Fagent-runtime.md) before selecting the runner.\n\n## Select the runner\n\nDiscover the installed agent's supported automation interface. Inspect its CLI help or server API before writing the adapter. A chat interface alone is not a webhook receiver. Load `$headless` for each CLI's non-interactive flags. Its `headless.py` launcher runs blocking reviews and streams no job events.\n\n| Agent | Available integration to check |\n| --- | --- |\n| Codex | A server-owned `codex exec` process with JSON event output or a configured SDK integration |\n| Claude Code | A server-owned `claude -p` process with structured output or Agent SDK integration |\n| Pi | A supervised `pi --mode rpc` process, SDK session, or separate print-mode invocation |\n| OpenCode | Its configured server\u002FSDK session API, including asynchronous prompts where supported |\n\nUse an existing authenticated webhook service if the user already has one. Read its actual URL, authentication scheme, request format, and success semantics. Otherwise implement the local server as the adapter to the selected runner. Record whether jobs start fresh sessions or resume a specific task session. Serialize requests to a shared session or checkout.\n\nIf no callable runner is available, prepare the UI and request contract, then report that starting agent jobs is blocked. Do not claim that a button can wake the current chat or survive a restart without a configured service that does so.\n\n## Define the job contract\n\nName the allowed actions and their JSON fields. Validate input at the server boundary. Pass data separately from fixed task instructions and use process argument arrays or structured SDK calls, never a shell command assembled from request text.\n\nReturn a job ID when the server accepts work. Report queued, running, completed, or failed from the runner's actual events and exit status. An accepted HTTP request does not prove that the agent started or completed the task. Use an idempotency key if requests can be retried. Persist pending jobs only when recovery is required, and mark interrupted jobs explicitly after restart.\n\n## Configure credentials\n\nUse the environment's secret-entry facility when available. Otherwise ask the user to enter credentials directly into a server-side secret store or restricted local configuration file. Never request secret values in chat. Follow the repository's managed-file rules for configuration.\n\nKeep provider credentials and upstream webhook keys out of browser code, responses, URLs, screenshots, and logs. Use the upstream service's documented authentication headers. Do not copy authentication headers from another provider.\n\n## Serve the UI\n\nServe the page and job endpoints from one origin. Keep the agent API behind this server. Authenticate job submission, enforce the allowed action list, and restrict browser origins. Set a short timeout for accepting a request and a separate job runtime limit. A slow agent must not hold a browser request open indefinitely.\n\nPrefer a localhost listener behind an existing Tailscale proxy. If using a direct tailnet listener, bind to the node's Tailscale address and confirm the interface. Do not expose an unauthenticated agent API on all network interfaces.\n\nUse an existing online Tailscale node. Inspect `tailscale status` and `tailscale ip -4`, then report the actual reachable URL. Follow the installed version's setup workflow if the node is not connected. Authentication belongs in the user's browser or secret-entry UI, not in chat. Do not install or reconfigure another node without need.\n\n## Prove the complete path\n\nUse a harmless allowed action. Click the real page, observe the accepted job ID, then inspect the runner's start event, completion event, and expected result. Verify that failed starts and malformed input produce useful errors. If recovery or retries are supported, prove that resubmission does not execute the same job twice.\n\nReport the UI URL, selected runner, where job status is visible, and what must remain running. A local server stops providing service when its host stops unless another configured host takes over.\n\nIntegration references: [Codex non-interactive mode](https:\u002F\u002Fdevelopers.openai.com\u002Fcodex\u002Fnoninteractive), [Claude Code programmatic use](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fheadless), [Pi RPC](https:\u002F\u002Fgithub.com\u002Fbadlogic\u002Fpi-mono\u002Fblob\u002Fmain\u002Fpackages\u002Fcoding-agent\u002Fdocs\u002Frpc.md), [OpenCode server](https:\u002F\u002Fopencode.ai\u002Fdocs\u002Fserver\u002F).\n",{"kind":24,"name":5,"role":25,"description":7},"dev","helper","main","c2b929d1a54305e77aad0296872e90e08d68a637","authoring\u002Fpoteto\u002Fmake-bot-ui",1,"https:\u002F\u002Fgithub.com\u002Fpascalandy\u002Fskills","https:\u002F\u002Fgithub.com\u002Fpascalandy","https:\u002F\u002Fblog.pascalandy.com","The question: how can I build the default one-stop place where entrepreneurs can start their mobile-first websites and blog? Psst psst… play-with-ghost.com","2026-11-15 03:34:59",{"source":8,"skillName":5,"commands":36,"prompt":46,"zipAvailable":47},[37,40,43],{"tool":38,"command":39},"npx","npx skills add pascalandy\u002Fskills --skill make-bot-ui",{"tool":41,"command":42},"bunx","bunx skills add pascalandy\u002Fskills --skill make-bot-ui",{"tool":44,"command":45},"pnpm","pnpm dlx skills add pascalandy\u002Fskills --skill make-bot-ui","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fpascalandy\u002Fskills\u002Fblob\u002Fmain\u002Fauthoring\u002Fpoteto\u002Fmake-bot-ui\u002FSKILL.md",true,[],[50,57,64,71,78,85],{"skillNo":51,"name":52,"nameZh":6,"description":53,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":19},"01M7DEA67FREYEGQJYTWD9D0GB","writing-great-skills","Use when creating, modifying, evaluating a skill.","https:\u002F\u002Fgithub.com\u002Fpascalandy\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-great-skills\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M7DEA3ZN566HSV6YJ8KPE824","writing-for-agents","Use when creating, editing, or reviewing a skill, AGENTS.md, CLAUDE.md, or another document agents read.","https:\u002F\u002Fgithub.com\u002Fpascalandy\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-for-agents\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":19},"01M7DEA0FRE0FEYV9XS0YTTBPJ","why","Use for questions about design rationale, the history of regressions or incidents, or the evidence behind thresholds and tradeoffs. Use `how` for runtime behavior. Do not use for plain commit or date lookups.","https:\u002F\u002Fgithub.com\u002Fpascalandy\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fwhy\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M7DE9XVZASJS9PC1A5BVMVHG","verify-video-archive","Use when validating the macOS or Linux archive workflow reached by `just convert-video`, including real media, prerequisites, terminal progress, source safety, locking, and recovery.","https:\u002F\u002Fgithub.com\u002Fpascalandy\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fverify-video-archive\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M7DE9VTG6ZEQH1M10J66T8MR","verify-transcript","Use when validating transcript CLI behavior, locating its verification features, or running the paid YouTube end-to-end check.","https:\u002F\u002Fgithub.com\u002Fpascalandy\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fverify-transcript\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":19},"01M7DE9RWQ6WNVG455HQFHVNFR","verify-skills","Use when verifying `just compile-skills`, `just remote-skills`, `just install-skills`, or `just skills-discover` in the skills repository, such as after changing their scripts or adding, renaming, or moving a skill.","https:\u002F\u002Fgithub.com\u002Fpascalandy\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fverify-skills\u002FSKILL.md",[],[]]