Agent-Skills.md

Agent Skills: Algolia Data Handling Review

'Implement Algolia data handling: record transforms, PII filtering before

UncategorizedID: jeremylongshore/claude-code-plugins/algolia-data-handling

Repository

jeremylongshore/claude-code-plugins
jeremylongshoreLicense: MIT
2,458348

Install this agent skill to your local

pnpm dlx add-skill https://github.com/jeremylongshore/claude-code-plugins-plus-skills/tree/HEAD/plugins/saas-packs/algolia-pack/skills/algolia-data-handling

Skill Files

Browse the full folder contents for algolia-data-handling.

Download Skill

Loading file tree…

plugins/saas-packs/algolia-pack/skills/algolia-data-handling/SKILL.md

Skill Metadata

Name
algolia-data-handling
Description
>-

Algolia Data Handling Review

Overview

This skill maps which data leaves the source system, how it becomes searchable records or events, and how correction and deletion propagate. It does not claim legal compliance; it produces technical evidence for the responsible privacy owner.

Prerequisites

  • A named repository, environment, and Algolia application or index in scope
  • The local lockfile and installed client types as implementation authority
  • A safe read-only query or explicitly disposable test target
  • Current first-party documentation for any provider behavior that affects the change

Tool Discipline

Use Read, Glob, and Grep to inspect local code, configuration names, tests, and dependency versions. Use WebFetch only for current official Algolia documentation. Use Write or Edit only after identifying the target files, constraints, and verification plan.

Current Contract

  • Index only fields required for retrieval, ranking, filtering, display, or approved analytics.
  • Treat attributesToRetrieve as response shaping, not as a substitute for excluding sensitive data from records.
  • Keep source-system identity mappings so corrections and deletions are reproducible.
  • Handle record deletion and Insights user-token deletion as separate surfaces with separate evidence.

Authentication

Use a restricted backend write key for record changes and the documented authorization for user-data deletion. Never include raw secrets or direct personal identifiers in logs or example events.

Instructions

  1. Inventory record fields, derived attributes, event fields, user tokens, environments, and downstream exports.
  2. Classify each field by purpose, sensitivity, source authority, and deletion requirement.
  3. Remove unnecessary fields before indexing and test that UI and ranking behavior still work.
  4. Implement idempotent correction and deletion paths with request IDs and target indices.
  5. Verify absence using bounded lookups and preserve evidence without retaining the deleted value.
  6. Document retention ownership, incident escalation, and gaps for privacy or legal review.

Approval Boundaries

Do not make legal conclusions, bulk-delete records, change retention policy, or expose protected values while verifying a request.

Output

Return the data-flow map, field inventory, minimization changes, deletion/correction procedure, verification evidence, unresolved provider retention questions, and named policy owner.

Error Handling

| Condition | Response | |---|---| | Identity mapping missing | Stop and reconcile the source identity before deletion. | | Deletion task incomplete | Retain the task ID and verify state before closing. | | Event token contains PII | Stop sending it and escalate remediation. | | Policy answer unavailable | Record the question for the privacy owner or provider. |

Examples

Use this compact input and expected handoff to calibrate scope and evidence quality.

Input:

request=privacy-123; surfaces=records,events; indices=customer_search

Expected handoff:

records=removed-and-verified; events=requested; legal-determination=not-made

Resources