Every premium and members-only resource for $9.99/month, as a no-commitment subscription. See the plans

PalateKit

Menu

Request a resource

Describe the resource you are missing in one sentence. We will email you when it is published.

0 / 280

Unlock PalateKit

Unlock everything

Plan

Paying means accepting our terms of sale.

Secure payment by Stripe. Receipts and invoices sent by email.

Welcome back

Sign in to install, download or save this resource. A free account is enough for free resources.

Search and recommendation planning for a marketplace v1.0.0

Use this resource to plan or diagnose search and recommendations for a two-sided marketplace on OpenSearch

New

  • 375 downloads
  • 159 likes
  • 143 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install marketplace-search-planning

With no option, the skill goes to ~/.agents/skills/; add --agent claude-code (or codex, cursor…) to target an agent.

First time: connect the CLI with

npx --yes palatekit@latest login

The agent frames a marketplace's search against 57 rules: user intent, surface architecture, index mapping, query understanding, retrieval strategy and ranking. It offers two playbooks, planning a system from scratch or diagnosing an existing one. You get a work plan ordered by impact.

Mostly picked by

Updated November 2, 2026

Card for the “Search and recommendation planning for a marketplace” skill

Details

Generates with
Your agent's model
Licence
MIT

Similar resources

View all

What you get

  • references/: cards on surface architecture and candidate generation
  • references/: cards on blending search and personalisation and zero-result fallbacks

Good fit for

  • You are launching a new search homefeed for a marketplace
  • Users find a marketplace's search results irrelevant

Not for

    Tags

    • marketplace
    • search
    • recommendation
    • opensearch
    • ranking

    About

    You say

    “Diagnose our rental search: results look relevant but conversion is dropping. Where should we start?”

    Card for the “Search and recommendation planning for a marketplace” skill

    Skill published by pproenca under the MIT license, delivered in its original version.

    Show more

    How to use it

    1. Install the skill in your agent's skills folder
    2. Ask your agent to plan or diagnose search and recommendations for a two-sided marketplace on OpenSearch
    3. Review and adjust the result

    Content preview

    Files

    • AGENTS.md16 KB
    • assets/templates/_template.md2.5 KB
    • gotchas.md1.5 KB
    • LICENSE1.1 KB
    • metadata.json3.2 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • README.md7.3 KB
    • references/_sections.md3.9 KB
    • references/arch-avoid-mono-stack-retrieval.md1.9 KB
    • references/arch-design-for-cold-start-from-day-one.md1.9 KB
    • references/arch-design-zero-result-fallback.md2.2 KB
    • references/arch-map-surface-to-retrieval-primitive.md1.8 KB
    • references/arch-route-surfaces-deliberately.md2.2 KB
    • references/arch-split-candidate-generation-from-ranking.md2.5 KB
    • references/blend-combine-search-and-personalisation-scores.md2.1 KB
    • references/blend-keep-hybrid-blending-explainable.md1.7 KB
    • references/blend-never-return-zero-results.md2 KB
    • references/blend-use-search-alone-for-specific-intent.md1.6 KB
    • references/index-design-mappings-conservatively.md1.9 KB
    • references/index-match-index-and-query-time-analyzers.md1.7 KB
    • references/index-separate-searchable-from-display-fields.md1.8 KB
    • references/index-stream-listing-updates-via-cdc.md2.1 KB
    • references/index-use-index-templates-for-consistency.md2.2 KB
    • references/index-use-keyword-and-text-as-multi-fields.md1.4 KB
    • references/index-use-language-analyzers-for-language-fields.md1.6 KB
    • references/intent-audit-live-query-logs-first.md1.9 KB
    • references/intent-distinguish-transactional-from-exploratory.md1.6 KB
    • references/intent-map-queries-to-intent-classes.md1.8 KB
    • references/intent-reject-one-search-for-everything.md1.9 KB
    • references/intent-separate-known-item-from-discovery.md1.8 KB
    • references/intent-treat-no-search-as-first-class-choice.md1.8 KB
    • references/measure-define-session-success-per-surface.md1.8 KB
    • references/measure-run-interleaving-as-cheap-ab-proxy.md1.7 KB
    • references/measure-track-ndcg-mrr-zero-result-rate.md1.6 KB
    • references/measure-track-reformulation-rate-as-failure-signal.md1.8 KB
    • references/measure-use-click-models-for-implicit-judgments.md2.1 KB
    • references/monitor-alert-on-decision-triggers.md2.2 KB
    • references/monitor-build-search-health-dashboard.md2 KB
    • references/monitor-log-every-query-with-full-context.md1.9 KB
    • references/monitor-run-weekly-search-quality-review.md1.8 KB
    • references/monitor-scrub-pii-from-query-logs.md2.4 KB
    • references/monitor-track-ranking-stability-churn.md1.9 KB
    • references/plan-audit-before-you-build.md1.9 KB
    • references/plan-build-golden-query-set-first.md1.8 KB
    • references/plan-find-bottleneck-before-optimising.md1.9 KB
    • references/plan-handoff-to-personalisation-skill.md2 KB
    • references/plan-maintain-a-decisions-log.md2 KB
    • references/plan-version-the-golden-set.md1.7 KB
    • references/playbooks/improving.md13.8 KB
    • references/playbooks/planning.md10.9 KB
    • references/query-build-autocomplete-on-separate-index.md1.5 KB
    • references/query-classify-before-routing.md1.9 KB
    • references/query-curate-synonyms-by-domain.md1.8 KB
    • references/query-normalise-before-anything-else.md1.8 KB
    • references/query-use-fuzzy-matching-for-typos.md1.3 KB
    • references/query-use-language-analyzers-for-stemming.md1.3 KB
    • references/rank-apply-diversity-at-rank-time.md1.8 KB
    • references/rank-deploy-ltr-only-after-golden-set-exists.md2 KB
    • references/rank-normalise-scores-across-retrieval-primitives.md2.2 KB
    • references/rank-tune-bm25-parameters-last.md1.5 KB
    • references/rank-use-function-score-for-business-signals.md1.9 KB
    • references/retrieve-choose-embedding-model-deliberately.md2.3 KB
    • references/retrieve-combine-bm25-and-knn-via-hybrid-search.md1.8 KB
    • references/retrieve-paginate-with-search-after.md1.4 KB
    • references/retrieve-run-expensive-signals-in-rescore.md1.7 KB
    • references/retrieve-use-bool-structure-deliberately.md1.6 KB
    • references/retrieve-use-filter-clauses-for-exact-matches.md1.5 KB
    • SKILL.md14.6 KB

    Origin and licence

    Listed unmodified

    Author
    pproenca
    Source
    github.com/pproenca/dot-skills/tree/cf93c57cac89d6fc3e4194686000411567f5caf3/skills/.experimental/marketplace-search-recsys-planning
    Licence
    MIT

    Notice

    « marketplace-search-recsys-planning » (pproenca, MIT) livré dans sa version d'origine : fiche en français et en anglais par PalateKit.

    User reviews

    –

    No ratings yet

    Loading reviews…

    Delete your review?

    Your rating and comment will be removed from the page.