| name | algolia-bootstrap |
| description | Stand up Algolia for a new engagement — applications per environment, indices, keys with least-privilege ACLs, settings-as-code repo layout, Insights wiring, and the Contentful indexer scaffold. Orchestrator that sequences across the algolia-* skills end-to-end. Use this skill on greenfield Composable DXP engagements when Algolia is being introduced; do NOT run on an existing Algolia integration with production records (that's a restructure, not a bootstrap — use algolia-audit + algolia-relevance-review instead). |
| type | skill |
| project | skills-library |
| plugin | algolia |
| aliases | ["algolia-bootstrap"] |
| tags | ["type/skill","plugin/algolia","topic/algolia","topic/bootstrap","scope/orchestrator"] |
| status | active |
Algolia bootstrap (orchestrator)
Use the algolia-agent to set up a new Algolia integration end-to-end. This is the greenfield runbook for a Composable DXP engagement.
Pick the engagement context first — Contentful + Next.js (default), commerce-on-Shopify, or other source. The choice calibrates Stage 4 onward.
The flow runs 8 stages, each delegating to the right skill:
-
Applications + plan — confirm the active plan tier (Build / Grow / Premium) and provision per-environment apps (slalom-{client}-prod, -staging, -dev). Premium-only features (NeuralSearch, AI Browse) are flagged for Stage 7. See references/slalom-context.md for the partner-portal sandbox path.
-
Index design — algolia-index-design decides per-type indices vs. federated, locale fan-out, replicas (default virtual), distinct strategy. Output: an attribute matrix per index, committed to the engagement's docs.
-
Records and source mapping — for Contentful sources, algolia-contentful-integration picks Marketplace app vs. custom Vercel-Function indexer. For other sources, see references/integrations-map.md. Output: a mapping module per content type.
-
Indexing pipeline — algolia-indexing-pipeline scaffolds the webhook handler, signature verification, atomic-reindex script, and scheduled drift-repair cron. Default tier 2 (Vercel Function + Contentful webhook).
-
Keys — algolia-api-keys-security mints scoped keys per workload (indexer, settings-deploy, browser search-only, MCP read). Stores them in the engagement's secret store. Documents the secured-API-key pattern if multi-tenant.
-
Settings, rules, synonyms — as code — establish algolia/ at repo root with *.settings.json, *.rules.json, *.synonyms.json per index. CI workflow deploys staging on merge, prod on manual trigger. See algolia-mcp-cli for the deploy script and references/cli-cheat-sheet.md for the CLI commands.
-
Front-end wiring — algolia-instantsearch-react for the SERP, algolia-autocomplete for the global header, algolia-search-client for any custom or RSC surface. Hand-off to software-engineering-frontend-developer for the design-system primitives.
-
Insights + analytics baseline — algolia-analytics-events writes the event taxonomy (event names, userToken policy, queryID attribution). Wire on day one. Defer Personalization, Recommend, NeuralSearch to a later A/B once events have flowed for ≥30 days.
Expected duration: a focused 1–2 day kickoff, plus 1–2 weeks of incremental wiring before the SERP goes live.
Don't skip Stage 8. Without Insights, Recommend and Personalization can't train, A/B tests can't measure, and Search Analytics flies blind.
Don't enable AI features (NeuralSearch, AI Browse, Personalization) during bootstrap. Capture them as a Phase 2 backlog item that runs after the events history is sufficient.
After bootstrap completes, the integration is ready to operate. The first algolia-audit should run ~30 days later to validate the events are flowing and surface any settings drift.
If the engagement already has Algolia in production with active records, this is a restructure, not a bootstrap. Use algolia-audit and algolia-relevance-review instead.
