{"id":7263,"plugin_id":"plugin_asdk_app_6a566052cde08191bbb62f836aba2c93","kind":"skill","collection_source":null,"comparison_source":null,"observed_at":"2026-09-30T22:50:14.553Z","digest":"fa17603c1e3ae62566ef46a28cd78dbe522428a1f109f1a4fe951aae9fe392ef","against":null,"payload":{"name":"build-an-agent","description":"Use when the user wants to build, configure, or publish an Outside Agent — a hosted AI agent that talks to end users over SMS, voice, and web. Drives the outside-agent MCP tools end to end.","included_files":[],"skill_md_contents":"---\nname: build-an-agent\ndescription: Use when the user wants to build, configure, or publish an Outside Agent — a hosted AI agent that talks to end users over SMS, voice, and web. Drives the outside-agent MCP tools end to end.\n---\n\n# Build an Outside Agent\n\nYou have the `outside-agent` MCP server connected. It lets you compose a hosted consumer\nAI agent for the signed-in creator and publish it to SMS, voice, and web.\n\n## Interview first, don't build blind\n\nA short brief always leaves design decisions unmade. Before publishing, walk the design\ntree WITH the creator:\n\n- The build tools (`createAgent`, `createConversationalAgent`, `setAuthoredConfig`,\n  `patchAuthoredConfig`) return a `design.openDecisions` array — the dimensions still\n  unpinned (purpose, audience, soul, capabilities, guidance, channels), each with the\n  question to ask AND a recommended answer.\n- Put those questions to the creator ONE AT A TIME, in order, each with your recommended\n  answer, and wait for their reply before the next.\n- If a decision is answerable by exploring (`getCatalogSummary` for what it can do,\n  `listConnectors` for an integration), explore and decide instead of asking.\n\n## Workflow\n\n1. `createAgent` (name it) — or `createConversationalAgent` for a plain persona agent in one call.\n2. `recordSourceRequirements` — capture the raw brief + each interview Q&A before you review.\n3. `setAuthoredConfig` / `patchAuthoredConfig` — write the whole config (persona, soul,\n   capabilities, guidance, channels).\n4. `seedKnowledgeText` — ground the agent in the creator's real content if it answers questions.\n5. `lintConfig` → `simulateBuild` → `reviewAgent` — validate and safety-review.\n6. `publishAgent` (or `publishAndDeploy`) — ship it. Confirm any `assumedDefaults` it returns\n   with the creator rather than treating publish as the end.\n\nPrefer `getCatalog` / `getCatalogSummary` to learn the real capability + connector catalog\nbefore configuring, so you author against what the platform actually supports.\n"},"changes":[],"summary":"First saved snapshot. No earlier version is available for comparison.","summary_kind":"deterministic","summary_metadata":{}}