[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-algolia-algolia-agent-studio":3,"mdc--c0k5m7-key":39,"related-repo-algolia-algolia-agent-studio":606,"related-org-algolia-algolia-agent-studio":701},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":34,"sourceUrl":37,"mdContent":38},"algolia-agent-studio","build conversational agents with Algolia Agent Studio","Product-specific Algolia Agent Studio implementation, validation, and optimization guidance. Use when planning, building, integrating, or auditing Agent Studio agents, AI-powered conversational experiences, LLM provider setup, Algolia Search tools, client-side tools, MCP tools, memory, prompting, conversations, turn context, caching, analytics, feedback, authentication, approved domains, guardrails, or tool security. Do NOT use for live Agent Studio configuration, dry runs, publish\u002Fdeploy actions, or config-as-code operations; use the official algobot-cli skill instead. Do NOT use for generic non-Algolia RAG or chatbot architecture unless Agent Studio is the target product.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},"algolia","Algolia","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Falgolia.png",[12,14,17,20],{"name":9,"slug":8,"type":13},"tag",{"name":15,"slug":16,"type":13},"AI","ai",{"name":18,"slug":19,"type":13},"Search","search",{"name":21,"slug":22,"type":13},"Agents","agents",8,"https:\u002F\u002Fgithub.com\u002Falgolia\u002Fskills","2026-08-01T06:06:28.033767","MIT",2,[29,30,8,31,32,19,33],"agent","ai-assistant","mcp","recommendations","skills",{"repoUrl":24,"stars":23,"forks":27,"topics":35,"description":36},[29,30,8,31,32,19,33],"Algolia skills for AI Agents","https:\u002F\u002Fgithub.com\u002Falgolia\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Falgolia-agent-studio","---\nname: algolia-agent-studio\ndescription: >\n  Product-specific Algolia Agent Studio implementation, validation, and optimization guidance. Use when planning, building, integrating, or auditing Agent Studio agents, AI-powered conversational experiences, LLM provider setup, Algolia Search tools, client-side tools, MCP tools, memory, prompting, conversations, turn context, caching, analytics, feedback, authentication, approved domains, guardrails, or tool security. Do NOT use for live Agent Studio configuration, dry runs, publish\u002Fdeploy actions, or config-as-code operations; use the official algobot-cli skill instead. Do NOT use for generic non-Algolia RAG or chatbot architecture unless Agent Studio is the target product.\nlicense: MIT\nmetadata:\n  author: algolia\n  version: \"0.4\"\n---\n\n# Algolia Agent Studio\n\nUse this skill for Agent Studio work after confirming the product goal and the quality of the underlying Algolia data, events, and search configuration. Agent Studio experiences are only as good as the tools, records, guardrails, and measurement loops they can rely on.\n\n## Customer-Facing Standard\n\n- Define the agent contract before setup: purpose, audience, tools, allowed actions, guardrails, escalation, and measurement.\n- Treat data, events, permissions, and tool security as prerequisites, not follow-up polish.\n- Use public `academy.algolia.com` for learning alignment and public `algolia.com\u002Fdoc` for current implementation guidance when source-backed context is needed.\n- When an Academy metadata reference pack is available, use only its `title`, `url`, `course`, `module`, `learning_objectives`, and `updated_at` fields for structure. If it is stale or no match exists, fall back to live Academy\u002Fdocs lookup. Do not treat cached metadata as course content or implementation authority.\n- Do not require custom Academy\u002Fdocs access; use customer-provided sources only as optional context.\n- When live Algolia data, analytics, index inspection, settings changes, or account actions are needed, use Algolia MCP, the Algolia CLI, or official Algolia skills for the live operation, then apply this skill to interpret results and validate the customer-ready implementation path.\n- Produce a launch recommendation: launch, limited rollout, fix first, or do not start.\n- Start the first agent with one narrow, high-intent job and the smallest tool\u002Fdata surface that can solve it; breadth comes after evidence.\n\n## Official Companion Skills\n\n- Use official `algobot-cli` for Agent Studio setup, agent config-as-code, tools, memory, conversations, dry runs, and deploy\u002Fpublish workflows.\n- Use official `algolia-mcp` for live search, analytics, recommendations, and index inspection that inform the agent's tools.\n- Use official `algolia-cli` for index, settings, rules, synonyms, or record changes the agent depends on.\n- Use this skill after the official tool call to produce the agent contract, readiness gates, security checks, event\u002Ffeedback plan, and launch recommendation.\n\n## Source-Of-Truth Rules\n\n- Use official `algobot-cli` and current Agent Studio docs for commands, config fields, provider behavior, completions endpoints, tool configuration, memory, analytics, feedback, and deploy\u002Fpublish steps.\n- Do not invent Agent Studio capabilities, endpoint URLs, tool permissions, or memory behavior.\n- If the official tool or docs cannot confirm a behavior, label it as an assumption and recommend a dry run or dashboard verification.\n\n## Workflow\n\n1. Read `references\u002Fagent-studio-guide.md` before implementation or review.\n2. Read `references\u002Fexample-output.md` when producing an agent contract or customer-facing launch recommendation.\n3. Start with `$algolia-discovery-planning` if business goal, audience, channel, or success metric is unclear.\n4. Use `$algolia-data-modeling` and `$algolia-index-configuration` when the agent will retrieve Algolia records or answer from search results.\n5. Use `$algolia-events-insights` when the experience needs analytics, feedback, personalization, Recommend, NeuralSearch optimization, or conversion measurement.\n6. Use `$algolia-release-qa` before launch, especially for guardrails, approved domains, authentication, tool security, and event coverage.\n7. Map the agent's room: instructions, tools, retrieval, data access, memory\u002Fcontext, safety, provider, and entry point. Diagnose behavior against that map before blaming the model.\n8. Teach the setup as a lifecycle: create and configure, preview and test, publish and integrate, then refine with Conversations and Analytics.\n\n## Questions To Ask\n\nAsk the smallest useful subset before building:\n\n- What job should the agent perform: discovery, product advice, support, shopping assistant, internal operations, merchandising, or content navigation?\n- Which entry point will users see first: search bar AI mode, autocomplete, side panel, InstantSearch Chat Widget, embedded assistant, or prompt suggestions?\n- Which Algolia indices, tools, or external systems may the agent use?\n- Which user attributes, session state, turn context, or memory are allowed?\n- Which actions are read-only versus write\u002Faction-taking?\n- What should the agent refuse, escalate, or hand off?\n- What LLM provider, model, region, latency, and cost constraints apply?\n- What analytics, feedback, event, or conversion signals define success?\n- What domains, auth model, and API-key boundaries are required?\n- Which one high-intent user task should the first rollout solve, and what requests are deliberately out of scope?\n- Which entry point best matches that task: search AI mode, chat widget, side panel, embedded assistant, or prompt suggestion?\n- For each enabled tool, what triggers it, what data\u002Farguments may it use, what result does the user see, and what happens on timeout or refusal?\n\n## Implementation Standards\n\n- Define the agent contract before code: purpose, audience, tools, allowed actions, guardrails, escalation paths, and measurement.\n- Treat Quick Setup or templates as starting points, not launch readiness. Review instructions, Search tool settings, memory, prompt suggestions, guardrails, rate limits, cost controls, approved domains, and integration snippet before recommending publish.\n- Treat Algolia Search tools as product APIs. Confirm index readiness, filters, searchable attributes, ranking, secured keys, and returned fields.\n- Never expose admin keys, unrestricted actions, internal-only data, or unreviewed external tools to a client-facing agent.\n- Keep prompts grounded in business policy and available tools. Avoid prompts that imply access to data or actions the agent cannot actually use.\n- Instrument feedback and conversion loops early; AI product tuning without events becomes anecdotal.\n- Validate failure modes: no results, ambiguous user intent, unavailable items, policy refusal, tool errors, auth failures, and slow LLM\u002Ftool responses.\n- Use Conversations for qualitative review before relying on aggregate analytics. Look for unclear instructions, bad follow-up behavior, weak retrieval, inappropriate fallback, and guardrail misses.\n- Use Analytics as a feedback loop, not a vanity dashboard. Let usage, token\u002Fcost patterns, search usage, feedback, and outcome signals drive configuration changes.\n- Define each enabled tool as a contract: trigger, allowed parameters, access scope, confirmation requirement, user-visible result, failure response, and measurement event.\n- Treat memory as intentional product behavior. Enable it for authenticated, multi-turn value; document identity, retention, consent, and what must never be retained. Do not assume dashboard enablement alone creates safe end-to-end memory.\n- Choose the entry point by user moment. Search-side AI helps query refinement and retrieval; a conversational surface supports multi-turn help; prompt suggestions should lead only to tasks the agent can actually complete.\n\n## Anti-Patterns\n\n- Starting with prompts before defining tools, permissions, data boundaries, and success measures.\n- Giving a client-facing agent write-capable tools before a read-only rollout has been validated.\n- Relying on prompt instructions as the only guardrail for sensitive data or consequential actions.\n- Shipping without approved domains, authentication review, fallback\u002Fhandoff behavior, and feedback measurement.\n- Treating Agent Studio as separate from search relevance, event quality, or index permissions.\n- Publishing immediately after the preview gives one good answer.\n- Treating Conversations as support logs only; they are the fastest way to identify bad instructions, weak retrieval, missing data, and fallback failures.\n- Expanding a first agent into a broad assistant before one high-intent, read-only workflow has reliable retrieval, tool use, fallback, and measurement.\n- Blaming the selected LLM before checking scope, instructions, tools, retrieval, data access, memory\u002Fcontext, safety, and integration state.\n\n## Academy And Customer Education Alignment\n\nWhen source-backed guidance is needed, search public Academy sources for Agent Studio and AI-readiness learning objectives and public Algolia docs for setup, tools, events, security, and current product behavior. Use source guidance procedurally, not as copied documentation. Map the request to a customer maturity level and one or more use-case bundles, then provide customer-facing setup, validation, and optimization guidance.\n\n## Maturity Behavior\n\n- Beginner implementation: define the agent job, audience, data sources, and simplest safe read-only flow before tool expansion.\n- Production readiness: validate authentication, approved domains, tool boundaries, API keys, fallback behavior, event coverage, handoff paths, and the agent-room troubleshooting trace.\n- Optimization: add feedback loops, conversation analytics, tool-result quality review, and measured prompt\u002Ftool iteration.\n- AI readiness: require clean searchable data, reliable events, permission-aware tools, security review, and a clear launch or fix-first recommendation.\n\n## Output Contract\n\nFor implementation, return an agent contract, agent-room map, tool contracts, entry-point choice, setup steps, integration code or configuration, event\u002Ffeedback plan, security notes, QA cases, and a post-launch refinement loop. For audits, lead with launch blockers and the data\u002Fevent\u002Ftool dependencies that must be fixed first.\n",{"data":40,"body":43},{"name":4,"description":6,"license":26,"metadata":41},{"author":8,"version":42},"0.4",{"type":44,"children":45},"root",[46,54,60,67,174,180,225,231,255,261,358,364,369,432,438,501,507,555,561,566,572,595,601],{"type":47,"tag":48,"props":49,"children":50},"element","h1",{"id":4},[51],{"type":52,"value":53},"text","Algolia Agent Studio",{"type":47,"tag":55,"props":56,"children":57},"p",{},[58],{"type":52,"value":59},"Use this skill for Agent Studio work after confirming the product goal and the quality of the underlying Algolia data, events, and search configuration. Agent Studio experiences are only as good as the tools, records, guardrails, and measurement loops they can rely on.",{"type":47,"tag":61,"props":62,"children":64},"h2",{"id":63},"customer-facing-standard",[65],{"type":52,"value":66},"Customer-Facing Standard",{"type":47,"tag":68,"props":69,"children":70},"ul",{},[71,77,82,104,154,159,164,169],{"type":47,"tag":72,"props":73,"children":74},"li",{},[75],{"type":52,"value":76},"Define the agent contract before setup: purpose, audience, tools, allowed actions, guardrails, escalation, and measurement.",{"type":47,"tag":72,"props":78,"children":79},{},[80],{"type":52,"value":81},"Treat data, events, permissions, and tool security as prerequisites, not follow-up polish.",{"type":47,"tag":72,"props":83,"children":84},{},[85,87,94,96,102],{"type":52,"value":86},"Use public ",{"type":47,"tag":88,"props":89,"children":91},"code",{"className":90},[],[92],{"type":52,"value":93},"academy.algolia.com",{"type":52,"value":95}," for learning alignment and public ",{"type":47,"tag":88,"props":97,"children":99},{"className":98},[],[100],{"type":52,"value":101},"algolia.com\u002Fdoc",{"type":52,"value":103}," for current implementation guidance when source-backed context is needed.",{"type":47,"tag":72,"props":105,"children":106},{},[107,109,115,117,123,124,130,131,137,138,144,146,152],{"type":52,"value":108},"When an Academy metadata reference pack is available, use only its ",{"type":47,"tag":88,"props":110,"children":112},{"className":111},[],[113],{"type":52,"value":114},"title",{"type":52,"value":116},", ",{"type":47,"tag":88,"props":118,"children":120},{"className":119},[],[121],{"type":52,"value":122},"url",{"type":52,"value":116},{"type":47,"tag":88,"props":125,"children":127},{"className":126},[],[128],{"type":52,"value":129},"course",{"type":52,"value":116},{"type":47,"tag":88,"props":132,"children":134},{"className":133},[],[135],{"type":52,"value":136},"module",{"type":52,"value":116},{"type":47,"tag":88,"props":139,"children":141},{"className":140},[],[142],{"type":52,"value":143},"learning_objectives",{"type":52,"value":145},", and ",{"type":47,"tag":88,"props":147,"children":149},{"className":148},[],[150],{"type":52,"value":151},"updated_at",{"type":52,"value":153}," fields for structure. If it is stale or no match exists, fall back to live Academy\u002Fdocs lookup. Do not treat cached metadata as course content or implementation authority.",{"type":47,"tag":72,"props":155,"children":156},{},[157],{"type":52,"value":158},"Do not require custom Academy\u002Fdocs access; use customer-provided sources only as optional context.",{"type":47,"tag":72,"props":160,"children":161},{},[162],{"type":52,"value":163},"When live Algolia data, analytics, index inspection, settings changes, or account actions are needed, use Algolia MCP, the Algolia CLI, or official Algolia skills for the live operation, then apply this skill to interpret results and validate the customer-ready implementation path.",{"type":47,"tag":72,"props":165,"children":166},{},[167],{"type":52,"value":168},"Produce a launch recommendation: launch, limited rollout, fix first, or do not start.",{"type":47,"tag":72,"props":170,"children":171},{},[172],{"type":52,"value":173},"Start the first agent with one narrow, high-intent job and the smallest tool\u002Fdata surface that can solve it; breadth comes after evidence.",{"type":47,"tag":61,"props":175,"children":177},{"id":176},"official-companion-skills",[178],{"type":52,"value":179},"Official Companion Skills",{"type":47,"tag":68,"props":181,"children":182},{},[183,196,208,220],{"type":47,"tag":72,"props":184,"children":185},{},[186,188,194],{"type":52,"value":187},"Use official ",{"type":47,"tag":88,"props":189,"children":191},{"className":190},[],[192],{"type":52,"value":193},"algobot-cli",{"type":52,"value":195}," for Agent Studio setup, agent config-as-code, tools, memory, conversations, dry runs, and deploy\u002Fpublish workflows.",{"type":47,"tag":72,"props":197,"children":198},{},[199,200,206],{"type":52,"value":187},{"type":47,"tag":88,"props":201,"children":203},{"className":202},[],[204],{"type":52,"value":205},"algolia-mcp",{"type":52,"value":207}," for live search, analytics, recommendations, and index inspection that inform the agent's tools.",{"type":47,"tag":72,"props":209,"children":210},{},[211,212,218],{"type":52,"value":187},{"type":47,"tag":88,"props":213,"children":215},{"className":214},[],[216],{"type":52,"value":217},"algolia-cli",{"type":52,"value":219}," for index, settings, rules, synonyms, or record changes the agent depends on.",{"type":47,"tag":72,"props":221,"children":222},{},[223],{"type":52,"value":224},"Use this skill after the official tool call to produce the agent contract, readiness gates, security checks, event\u002Ffeedback plan, and launch recommendation.",{"type":47,"tag":61,"props":226,"children":228},{"id":227},"source-of-truth-rules",[229],{"type":52,"value":230},"Source-Of-Truth Rules",{"type":47,"tag":68,"props":232,"children":233},{},[234,245,250],{"type":47,"tag":72,"props":235,"children":236},{},[237,238,243],{"type":52,"value":187},{"type":47,"tag":88,"props":239,"children":241},{"className":240},[],[242],{"type":52,"value":193},{"type":52,"value":244}," and current Agent Studio docs for commands, config fields, provider behavior, completions endpoints, tool configuration, memory, analytics, feedback, and deploy\u002Fpublish steps.",{"type":47,"tag":72,"props":246,"children":247},{},[248],{"type":52,"value":249},"Do not invent Agent Studio capabilities, endpoint URLs, tool permissions, or memory behavior.",{"type":47,"tag":72,"props":251,"children":252},{},[253],{"type":52,"value":254},"If the official tool or docs cannot confirm a behavior, label it as an assumption and recommend a dry run or dashboard verification.",{"type":47,"tag":61,"props":256,"children":258},{"id":257},"workflow",[259],{"type":52,"value":260},"Workflow",{"type":47,"tag":262,"props":263,"children":264},"ol",{},[265,278,290,303,324,336,348,353],{"type":47,"tag":72,"props":266,"children":267},{},[268,270,276],{"type":52,"value":269},"Read ",{"type":47,"tag":88,"props":271,"children":273},{"className":272},[],[274],{"type":52,"value":275},"references\u002Fagent-studio-guide.md",{"type":52,"value":277}," before implementation or review.",{"type":47,"tag":72,"props":279,"children":280},{},[281,282,288],{"type":52,"value":269},{"type":47,"tag":88,"props":283,"children":285},{"className":284},[],[286],{"type":52,"value":287},"references\u002Fexample-output.md",{"type":52,"value":289}," when producing an agent contract or customer-facing launch recommendation.",{"type":47,"tag":72,"props":291,"children":292},{},[293,295,301],{"type":52,"value":294},"Start with ",{"type":47,"tag":88,"props":296,"children":298},{"className":297},[],[299],{"type":52,"value":300},"$algolia-discovery-planning",{"type":52,"value":302}," if business goal, audience, channel, or success metric is unclear.",{"type":47,"tag":72,"props":304,"children":305},{},[306,308,314,316,322],{"type":52,"value":307},"Use ",{"type":47,"tag":88,"props":309,"children":311},{"className":310},[],[312],{"type":52,"value":313},"$algolia-data-modeling",{"type":52,"value":315}," and ",{"type":47,"tag":88,"props":317,"children":319},{"className":318},[],[320],{"type":52,"value":321},"$algolia-index-configuration",{"type":52,"value":323}," when the agent will retrieve Algolia records or answer from search results.",{"type":47,"tag":72,"props":325,"children":326},{},[327,328,334],{"type":52,"value":307},{"type":47,"tag":88,"props":329,"children":331},{"className":330},[],[332],{"type":52,"value":333},"$algolia-events-insights",{"type":52,"value":335}," when the experience needs analytics, feedback, personalization, Recommend, NeuralSearch optimization, or conversion measurement.",{"type":47,"tag":72,"props":337,"children":338},{},[339,340,346],{"type":52,"value":307},{"type":47,"tag":88,"props":341,"children":343},{"className":342},[],[344],{"type":52,"value":345},"$algolia-release-qa",{"type":52,"value":347}," before launch, especially for guardrails, approved domains, authentication, tool security, and event coverage.",{"type":47,"tag":72,"props":349,"children":350},{},[351],{"type":52,"value":352},"Map the agent's room: instructions, tools, retrieval, data access, memory\u002Fcontext, safety, provider, and entry point. Diagnose behavior against that map before blaming the model.",{"type":47,"tag":72,"props":354,"children":355},{},[356],{"type":52,"value":357},"Teach the setup as a lifecycle: create and configure, preview and test, publish and integrate, then refine with Conversations and Analytics.",{"type":47,"tag":61,"props":359,"children":361},{"id":360},"questions-to-ask",[362],{"type":52,"value":363},"Questions To Ask",{"type":47,"tag":55,"props":365,"children":366},{},[367],{"type":52,"value":368},"Ask the smallest useful subset before building:",{"type":47,"tag":68,"props":370,"children":371},{},[372,377,382,387,392,397,402,407,412,417,422,427],{"type":47,"tag":72,"props":373,"children":374},{},[375],{"type":52,"value":376},"What job should the agent perform: discovery, product advice, support, shopping assistant, internal operations, merchandising, or content navigation?",{"type":47,"tag":72,"props":378,"children":379},{},[380],{"type":52,"value":381},"Which entry point will users see first: search bar AI mode, autocomplete, side panel, InstantSearch Chat Widget, embedded assistant, or prompt suggestions?",{"type":47,"tag":72,"props":383,"children":384},{},[385],{"type":52,"value":386},"Which Algolia indices, tools, or external systems may the agent use?",{"type":47,"tag":72,"props":388,"children":389},{},[390],{"type":52,"value":391},"Which user attributes, session state, turn context, or memory are allowed?",{"type":47,"tag":72,"props":393,"children":394},{},[395],{"type":52,"value":396},"Which actions are read-only versus write\u002Faction-taking?",{"type":47,"tag":72,"props":398,"children":399},{},[400],{"type":52,"value":401},"What should the agent refuse, escalate, or hand off?",{"type":47,"tag":72,"props":403,"children":404},{},[405],{"type":52,"value":406},"What LLM provider, model, region, latency, and cost constraints apply?",{"type":47,"tag":72,"props":408,"children":409},{},[410],{"type":52,"value":411},"What analytics, feedback, event, or conversion signals define success?",{"type":47,"tag":72,"props":413,"children":414},{},[415],{"type":52,"value":416},"What domains, auth model, and API-key boundaries are required?",{"type":47,"tag":72,"props":418,"children":419},{},[420],{"type":52,"value":421},"Which one high-intent user task should the first rollout solve, and what requests are deliberately out of scope?",{"type":47,"tag":72,"props":423,"children":424},{},[425],{"type":52,"value":426},"Which entry point best matches that task: search AI mode, chat widget, side panel, embedded assistant, or prompt suggestion?",{"type":47,"tag":72,"props":428,"children":429},{},[430],{"type":52,"value":431},"For each enabled tool, what triggers it, what data\u002Farguments may it use, what result does the user see, and what happens on timeout or refusal?",{"type":47,"tag":61,"props":433,"children":435},{"id":434},"implementation-standards",[436],{"type":52,"value":437},"Implementation Standards",{"type":47,"tag":68,"props":439,"children":440},{},[441,446,451,456,461,466,471,476,481,486,491,496],{"type":47,"tag":72,"props":442,"children":443},{},[444],{"type":52,"value":445},"Define the agent contract before code: purpose, audience, tools, allowed actions, guardrails, escalation paths, and measurement.",{"type":47,"tag":72,"props":447,"children":448},{},[449],{"type":52,"value":450},"Treat Quick Setup or templates as starting points, not launch readiness. Review instructions, Search tool settings, memory, prompt suggestions, guardrails, rate limits, cost controls, approved domains, and integration snippet before recommending publish.",{"type":47,"tag":72,"props":452,"children":453},{},[454],{"type":52,"value":455},"Treat Algolia Search tools as product APIs. Confirm index readiness, filters, searchable attributes, ranking, secured keys, and returned fields.",{"type":47,"tag":72,"props":457,"children":458},{},[459],{"type":52,"value":460},"Never expose admin keys, unrestricted actions, internal-only data, or unreviewed external tools to a client-facing agent.",{"type":47,"tag":72,"props":462,"children":463},{},[464],{"type":52,"value":465},"Keep prompts grounded in business policy and available tools. Avoid prompts that imply access to data or actions the agent cannot actually use.",{"type":47,"tag":72,"props":467,"children":468},{},[469],{"type":52,"value":470},"Instrument feedback and conversion loops early; AI product tuning without events becomes anecdotal.",{"type":47,"tag":72,"props":472,"children":473},{},[474],{"type":52,"value":475},"Validate failure modes: no results, ambiguous user intent, unavailable items, policy refusal, tool errors, auth failures, and slow LLM\u002Ftool responses.",{"type":47,"tag":72,"props":477,"children":478},{},[479],{"type":52,"value":480},"Use Conversations for qualitative review before relying on aggregate analytics. Look for unclear instructions, bad follow-up behavior, weak retrieval, inappropriate fallback, and guardrail misses.",{"type":47,"tag":72,"props":482,"children":483},{},[484],{"type":52,"value":485},"Use Analytics as a feedback loop, not a vanity dashboard. Let usage, token\u002Fcost patterns, search usage, feedback, and outcome signals drive configuration changes.",{"type":47,"tag":72,"props":487,"children":488},{},[489],{"type":52,"value":490},"Define each enabled tool as a contract: trigger, allowed parameters, access scope, confirmation requirement, user-visible result, failure response, and measurement event.",{"type":47,"tag":72,"props":492,"children":493},{},[494],{"type":52,"value":495},"Treat memory as intentional product behavior. Enable it for authenticated, multi-turn value; document identity, retention, consent, and what must never be retained. Do not assume dashboard enablement alone creates safe end-to-end memory.",{"type":47,"tag":72,"props":497,"children":498},{},[499],{"type":52,"value":500},"Choose the entry point by user moment. Search-side AI helps query refinement and retrieval; a conversational surface supports multi-turn help; prompt suggestions should lead only to tasks the agent can actually complete.",{"type":47,"tag":61,"props":502,"children":504},{"id":503},"anti-patterns",[505],{"type":52,"value":506},"Anti-Patterns",{"type":47,"tag":68,"props":508,"children":509},{},[510,515,520,525,530,535,540,545,550],{"type":47,"tag":72,"props":511,"children":512},{},[513],{"type":52,"value":514},"Starting with prompts before defining tools, permissions, data boundaries, and success measures.",{"type":47,"tag":72,"props":516,"children":517},{},[518],{"type":52,"value":519},"Giving a client-facing agent write-capable tools before a read-only rollout has been validated.",{"type":47,"tag":72,"props":521,"children":522},{},[523],{"type":52,"value":524},"Relying on prompt instructions as the only guardrail for sensitive data or consequential actions.",{"type":47,"tag":72,"props":526,"children":527},{},[528],{"type":52,"value":529},"Shipping without approved domains, authentication review, fallback\u002Fhandoff behavior, and feedback measurement.",{"type":47,"tag":72,"props":531,"children":532},{},[533],{"type":52,"value":534},"Treating Agent Studio as separate from search relevance, event quality, or index permissions.",{"type":47,"tag":72,"props":536,"children":537},{},[538],{"type":52,"value":539},"Publishing immediately after the preview gives one good answer.",{"type":47,"tag":72,"props":541,"children":542},{},[543],{"type":52,"value":544},"Treating Conversations as support logs only; they are the fastest way to identify bad instructions, weak retrieval, missing data, and fallback failures.",{"type":47,"tag":72,"props":546,"children":547},{},[548],{"type":52,"value":549},"Expanding a first agent into a broad assistant before one high-intent, read-only workflow has reliable retrieval, tool use, fallback, and measurement.",{"type":47,"tag":72,"props":551,"children":552},{},[553],{"type":52,"value":554},"Blaming the selected LLM before checking scope, instructions, tools, retrieval, data access, memory\u002Fcontext, safety, and integration state.",{"type":47,"tag":61,"props":556,"children":558},{"id":557},"academy-and-customer-education-alignment",[559],{"type":52,"value":560},"Academy And Customer Education Alignment",{"type":47,"tag":55,"props":562,"children":563},{},[564],{"type":52,"value":565},"When source-backed guidance is needed, search public Academy sources for Agent Studio and AI-readiness learning objectives and public Algolia docs for setup, tools, events, security, and current product behavior. Use source guidance procedurally, not as copied documentation. Map the request to a customer maturity level and one or more use-case bundles, then provide customer-facing setup, validation, and optimization guidance.",{"type":47,"tag":61,"props":567,"children":569},{"id":568},"maturity-behavior",[570],{"type":52,"value":571},"Maturity Behavior",{"type":47,"tag":68,"props":573,"children":574},{},[575,580,585,590],{"type":47,"tag":72,"props":576,"children":577},{},[578],{"type":52,"value":579},"Beginner implementation: define the agent job, audience, data sources, and simplest safe read-only flow before tool expansion.",{"type":47,"tag":72,"props":581,"children":582},{},[583],{"type":52,"value":584},"Production readiness: validate authentication, approved domains, tool boundaries, API keys, fallback behavior, event coverage, handoff paths, and the agent-room troubleshooting trace.",{"type":47,"tag":72,"props":586,"children":587},{},[588],{"type":52,"value":589},"Optimization: add feedback loops, conversation analytics, tool-result quality review, and measured prompt\u002Ftool iteration.",{"type":47,"tag":72,"props":591,"children":592},{},[593],{"type":52,"value":594},"AI readiness: require clean searchable data, reliable events, permission-aware tools, security review, and a clear launch or fix-first recommendation.",{"type":47,"tag":61,"props":596,"children":598},{"id":597},"output-contract",[599],{"type":52,"value":600},"Output Contract",{"type":47,"tag":55,"props":602,"children":603},{},[604],{"type":52,"value":605},"For implementation, return an agent contract, agent-room map, tool contracts, entry-point choice, setup steps, integration code or configuration, event\u002Ffeedback plan, security notes, QA cases, and a post-launch refinement loop. For audits, lead with launch blockers and the data\u002Fevent\u002Ftool dependencies that must be fixed first.",{"items":607,"total":700},[608,622,629,644,657,670,682],{"slug":193,"name":193,"fn":609,"description":610,"org":611,"tags":612,"stars":23,"repoUrl":24,"updatedAt":621},"build conversational AI with Algolia","Use for anything AI\u002Fagent\u002Fconversational built on Algolia: algobot CLI, Agent Studio, RAG systems, conversational product discovery, genAI content generation from search results (carousels, descriptions, headers), chatbots or recommendation agents using Algolia as retrieval, config-as-code workflows, multi-environment deploy (dev\u002Fstaging\u002Fprod), memory and personalization, MCP tool integrations, conversation history \u002F GDPR retention, or adding a chat widget alongside InstantSearch. Trigger on: \"algobot\", \"Agent Studio\", \"RAG with Algolia\", \"conversational experience\", \"AI agent\" + Algolia, \"genAI carousel\", \"chat widget\", or building AI features on top of Algolia search. Do NOT use for raw index ops (records, synonyms, settings) — use algolia-cli. Do NOT use for pure frontend search UI (InstantSearch, autocomplete) with no AI\u002Fagent layer.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[613,614,617,620],{"name":21,"slug":22,"type":13},{"name":615,"slug":616,"type":13},"Automation","automation",{"name":618,"slug":619,"type":13},"LLM","llm",{"name":18,"slug":19,"type":13},"2026-07-12T08:27:37.649724",{"slug":4,"name":4,"fn":5,"description":6,"org":623,"tags":624,"stars":23,"repoUrl":24,"updatedAt":25},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[625,626,627,628],{"name":21,"slug":22,"type":13},{"name":15,"slug":16,"type":13},{"name":9,"slug":8,"type":13},{"name":18,"slug":19,"type":13},{"slug":630,"name":630,"fn":631,"description":632,"org":633,"tags":634,"stars":23,"repoUrl":24,"updatedAt":643},"algolia-autocomplete","build Algolia autocomplete and query suggestions","Build and review Algolia Autocomplete and query suggestion experiences. Use when planning or implementing typeahead, query suggestions, recent searches, popular searches, federated autocomplete panels, product\u002Fcontent suggestions, detached mobile mode, plugins, keyboard navigation, insights events, or Autocomplete integration with InstantSearch. For net-new search or ecommerce builds, start with algolia-discovery-planning, which loads algolia-search-implementation so source strategy, data contract, and event taxonomy decisions are visible before autocomplete is marked ready. Do NOT use for full search results pages, browse pages, filters, pagination, or current refinements; use algolia-instantsearch-ui instead. Do NOT use for choosing between Algolia UI libraries; use algolia-ui-libraries. Do NOT use as the source of truth for current Autocomplete package APIs; use the official instantsearch skill and current docs alongside this customer-readiness skill.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[635,636,639,640],{"name":9,"slug":8,"type":13},{"name":637,"slug":638,"type":13},"Frontend","frontend",{"name":18,"slug":19,"type":13},{"name":641,"slug":642,"type":13},"UI Components","ui-components","2026-08-01T06:06:29.968123",{"slug":217,"name":217,"fn":645,"description":646,"org":647,"tags":648,"stars":23,"repoUrl":24,"updatedAt":656},"manage Algolia indices and accounts","Use this skill whenever a user wants to execute operations against Algolia indices or accounts — deleting records, copying\u002Fmigrating indices, backing up data, importing\u002Fexporting records, managing API keys, editing synonyms, configuring rules, changing settings like facets, clearing indices, or automating Algolia in CI\u002FCD pipelines. The key signal is that the user wants to *act on* their Algolia data or configuration (server-side \u002F backend \u002F admin operations), regardless of whether they mention \"CLI\" or \"command line.\" If someone names a specific Algolia index and wants to change, move, query, or manage it, use this skill. Do NOT use for frontend search UI work (InstantSearch, React components, autocomplete widgets), Algolia dashboard GUI questions, or evaluating Algolia vs. other providers.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[649,652,655],{"name":650,"slug":651,"type":13},"CLI","cli",{"name":653,"slug":654,"type":13},"Data Engineering","data-engineering",{"name":18,"slug":19,"type":13},"2026-07-12T08:27:35.085246",{"slug":658,"name":658,"fn":659,"description":660,"org":661,"tags":662,"stars":23,"repoUrl":24,"updatedAt":669},"algolia-crawler","crawl websites into Algolia indices","Use this skill whenever a user wants to crawl one or more web pages or a whole site and turn them into an Algolia index using the Algolia CLI — especially for RAG, AI search, semantic search, or Agent Studio retrieval. Triggers: \"index my website\u002Fdocs with Algolia\", \"set up the Algolia Crawler\", \"crawl this page for RAG\", \"scrape my site into Algolia\", \"build a knowledge base for my AI agent from these URLs\", writing or debugging a crawler recordExtractor, or handling JavaScript-rendered pages that won't index. It guides ingestion end-to-end with `algolia crawler` commands: inspect the page, write a RAG-optimized recordExtractor, validate with `algolia crawler test` BEFORE indexing, apply index settings explicitly, then reindex. Do NOT use for building the chatbot\u002Fagent layer itself (use algobot-cli), raw record\u002Fsynonym\u002Fsettings ops on an existing index (use algolia-cli), frontend search UI (use instantsearch), or read-only search\u002Fanalytics (use algolia-mcp).",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[663,664,665,666],{"name":615,"slug":616,"type":13},{"name":653,"slug":654,"type":13},{"name":18,"slug":19,"type":13},{"name":667,"slug":668,"type":13},"Web Scraping","web-scraping","2026-07-12T08:27:40.981109",{"slug":671,"name":671,"fn":672,"description":673,"org":674,"tags":675,"stars":23,"repoUrl":24,"updatedAt":681},"algolia-data-modeling","design Algolia data models and indices","Algolia data modeling and indexing guidance. Use before or alongside indexing records or building Algolia search UI for net-new search, browse, autocomplete, ecommerce, personalization, Dynamic Re-Ranking, recommendations, or analytics-aware implementations. Makes record shape, objectID, display fields, facets, ranking fields, and event attribution explicit decisions. Use for records, variants, SKUs, indices, replicas, searchable and faceting attributes, denormalization, merchandising fields, timestamps, inventory, event attribution, indexing pipelines, partial updates, secured data, multi-language or multi-region strategies, and migrations. Do NOT use for live imports, exports, record mutations, settings changes, or account actions; use algolia-cli or algolia-mcp. Do NOT use for frontend UI implementation; use algolia-instantsearch-ui, algolia-autocomplete, algolia-ui-libraries, or the official instantsearch skill.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[676,677,680],{"name":9,"slug":8,"type":13},{"name":678,"slug":679,"type":13},"Data Modeling","data-modeling",{"name":18,"slug":19,"type":13},"2026-08-01T06:06:01.500756",{"slug":683,"name":683,"fn":684,"description":685,"org":686,"tags":687,"stars":23,"repoUrl":24,"updatedAt":699},"algolia-discovery-planning","plan and audit Algolia search implementations","START HERE for any non-trivial Algolia work — building, adding, migrating, redesigning, auditing, or configuring search, browse, autocomplete, indexing, relevance, recommendations, personalization, merchandising, events, or analytics. Invoke this FIRST even when the task already seems scoped or the user names one specific feature (e.g. \"add InstantSearch\", \"build a storefront search\"): its job is to map the request to the full Algolia implementation lifecycle and load every companion skill each in-scope phase needs (algolia-search-implementation, algolia-data-modeling, algolia-index-configuration, algolia-ui-libraries, algolia-instantsearch-ui, algolia-autocomplete, algolia-events-insights, algolia-neuralsearch, algolia-agent-studio, algolia-release-qa) rather than jumping straight into a single skill. This skill plans and orchestrates; the focused companion skills and the official Algolia skills execute. Do NOT use for live account inspection or write actions; use algolia-mcp or algolia-cli for those.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[688,689,692,695,698],{"name":9,"slug":8,"type":13},{"name":690,"slug":691,"type":13},"Analytics","analytics",{"name":693,"slug":694,"type":13},"Configuration","configuration",{"name":696,"slug":697,"type":13},"Personalization","personalization",{"name":18,"slug":19,"type":13},"2026-08-01T06:06:03.021452",18,{"items":702,"total":823},[703,719,726,733,740,746,753,759,767,778,789,811],{"slug":704,"name":704,"fn":705,"description":706,"org":707,"tags":708,"stars":716,"repoUrl":717,"updatedAt":718},"algolia-docsearch-mcp","search developer documentation with Algolia","Use this skill when the user asks about public developer documentation, SDKs, APIs, libraries, frameworks, setup, configuration, or code examples. Fetch current docs from Algolia DocSearch MCP and cite source URLs.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[709,710,713,715],{"name":9,"slug":8,"type":13},{"name":711,"slug":712,"type":13},"Documentation","documentation",{"name":714,"slug":31,"type":13},"MCP",{"name":18,"slug":19,"type":13},4365,"https:\u002F\u002Fgithub.com\u002Falgolia\u002Fdocsearch","2026-08-01T06:06:11.572314",{"slug":193,"name":193,"fn":609,"description":610,"org":720,"tags":721,"stars":23,"repoUrl":24,"updatedAt":621},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[722,723,724,725],{"name":21,"slug":22,"type":13},{"name":615,"slug":616,"type":13},{"name":618,"slug":619,"type":13},{"name":18,"slug":19,"type":13},{"slug":4,"name":4,"fn":5,"description":6,"org":727,"tags":728,"stars":23,"repoUrl":24,"updatedAt":25},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[729,730,731,732],{"name":21,"slug":22,"type":13},{"name":15,"slug":16,"type":13},{"name":9,"slug":8,"type":13},{"name":18,"slug":19,"type":13},{"slug":630,"name":630,"fn":631,"description":632,"org":734,"tags":735,"stars":23,"repoUrl":24,"updatedAt":643},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[736,737,738,739],{"name":9,"slug":8,"type":13},{"name":637,"slug":638,"type":13},{"name":18,"slug":19,"type":13},{"name":641,"slug":642,"type":13},{"slug":217,"name":217,"fn":645,"description":646,"org":741,"tags":742,"stars":23,"repoUrl":24,"updatedAt":656},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[743,744,745],{"name":650,"slug":651,"type":13},{"name":653,"slug":654,"type":13},{"name":18,"slug":19,"type":13},{"slug":658,"name":658,"fn":659,"description":660,"org":747,"tags":748,"stars":23,"repoUrl":24,"updatedAt":669},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[749,750,751,752],{"name":615,"slug":616,"type":13},{"name":653,"slug":654,"type":13},{"name":18,"slug":19,"type":13},{"name":667,"slug":668,"type":13},{"slug":671,"name":671,"fn":672,"description":673,"org":754,"tags":755,"stars":23,"repoUrl":24,"updatedAt":681},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[756,757,758],{"name":9,"slug":8,"type":13},{"name":678,"slug":679,"type":13},{"name":18,"slug":19,"type":13},{"slug":683,"name":683,"fn":684,"description":685,"org":760,"tags":761,"stars":23,"repoUrl":24,"updatedAt":699},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[762,763,764,765,766],{"name":9,"slug":8,"type":13},{"name":690,"slug":691,"type":13},{"name":693,"slug":694,"type":13},{"name":696,"slug":697,"type":13},{"name":18,"slug":19,"type":13},{"slug":768,"name":768,"fn":769,"description":770,"org":771,"tags":772,"stars":23,"repoUrl":24,"updatedAt":777},"algolia-events-insights","instrument Algolia events for analytics","Algolia event instrumentation guidance for Insights, analytics, personalization, Dynamic Re-Ranking, Recommend, and merchandising feedback loops. Use for search, autocomplete, browse, ecommerce, personalization, recommendations, or analytics instrumentation. Makes event decisions explicit before an Algolia UI is considered ready. Use when implementing or auditing clickedObjectIDsAfterSearch, convertedObjectIDsAfterSearch, viewedObjectIDs, addedToCartObjectIDsAfterSearch, purchasedObjectIDsAfterSearch, userToken, queryID, eventName, eventSubtype, or frontend\u002Fbackend event pipelines. Do NOT use for live analytics retrieval, top-query inspection, or account-aware diagnostics; use algolia-mcp. Do NOT use for framework-specific InstantSearch or Autocomplete APIs; use the official instantsearch skill alongside this planning and validation skill.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[773,774,775,776],{"name":9,"slug":8,"type":13},{"name":690,"slug":691,"type":13},{"name":696,"slug":697,"type":13},{"name":18,"slug":19,"type":13},"2026-08-01T06:06:02.009712",{"slug":779,"name":779,"fn":780,"description":781,"org":782,"tags":783,"stars":23,"repoUrl":24,"updatedAt":788},"algolia-index-configuration","configure Algolia index settings and relevance","Algolia index settings and relevance configuration guidance. Use when configuring searchableAttributes, attributesForFaceting, customRanking, ranking, replicas, virtual replicas, rules, synonyms, typo tolerance, distinct, filters, optional filters, merchandising, browse\u002Fcategory relevance, or A\u002FB-testable relevance changes. Do NOT use for live settings writes, backups, copies, or operational account tasks; use algolia-cli or algolia-mcp instead. Do NOT use for record-shape or variant strategy; use algolia-data-modeling instead.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[784,785,786,787],{"name":9,"slug":8,"type":13},{"name":693,"slug":694,"type":13},{"name":678,"slug":679,"type":13},{"name":18,"slug":19,"type":13},"2026-08-01T06:06:34.19163",{"slug":790,"name":790,"fn":791,"description":792,"org":793,"tags":794,"stars":23,"repoUrl":24,"updatedAt":810},"algolia-instantsearch-ui","build Algolia InstantSearch frontend experiences","Build and review Algolia InstantSearch experiences in JavaScript, React, Vue, Angular (via InstantSearch.js; Angular InstantSearch is deprecated), or compatible frontend stacks. Use when planning or reviewing search results pages, browse\u002Fcategory pages, routing, widgets, filters, facets, sort-by, pagination, infinite hits, current refinements, insights middleware, SSR, or UI-state synchronization. For net-new search UI builds, start with algolia-discovery-planning, which loads algolia-search-implementation so data contract and event taxonomy decisions are visible before UI is marked ready. Do NOT use for autocomplete\u002Ftypeahead experiences before the user commits to a results page; use algolia-autocomplete instead. Do NOT use for choosing between Algolia UI libraries; use algolia-ui-libraries. Do NOT use as the source of truth for current framework APIs or code-level implementation details; use the official instantsearch skill and current docs alongside this customer-readiness skill.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[795,796,799,800,803,806,807],{"name":9,"slug":8,"type":13},{"name":797,"slug":798,"type":13},"Angular","angular",{"name":637,"slug":638,"type":13},{"name":801,"slug":802,"type":13},"JavaScript","javascript",{"name":804,"slug":805,"type":13},"React","react",{"name":18,"slug":19,"type":13},{"name":808,"slug":809,"type":13},"Vue","vue","2026-08-01T06:06:33.66513",{"slug":205,"name":205,"fn":812,"description":813,"org":814,"tags":815,"stars":23,"repoUrl":24,"updatedAt":822},"search Algolia indices and retrieve analytics","Search Algolia indices via the Algolia MCP server, retrieve analytics (top searches, no-result rates, click positions, user counts), and get product recommendations (bought-together, related, trending). Triggers on search, indexing, analytics, Algolia, recommendations, MCP.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[816,817,820,821],{"name":690,"slug":691,"type":13},{"name":818,"slug":819,"type":13},"API Development","api-development",{"name":714,"slug":31,"type":13},{"name":18,"slug":19,"type":13},"2026-07-12T08:27:36.376387",19]