AgentSkillsCN

opik-integrations-auditor

对Python SDK、TypeScript SDK以及OTEL/API后端中的Opik集成进行审计、对比与文档化。在新增集成、审查现有集成，或生成首版集成文档并开展差距分析时均可使用。

SKILL.md
--- frontmatter
name: opik-integrations-auditor
description: Audit, compare, and document Opik integrations across Python SDK, TypeScript SDK, and OTEL/API backend. Use when adding a new integration, reviewing an existing one, or generating first-pass integration docs and gap analysis.
metadata:
  internal: true

Opik Integrations Auditor

Purpose

Audit, compare, and document Opik integrations across Python, TypeScript, and backend OTEL/API surfaces with a repeatable evidence-first workflow.

When to use

Use this skill when the task involves:

  • •reviewing existing Opik integrations for behavior, coverage, or documentation quality
  • •designing or implementing a new integration
  • •producing integration pattern docs across Python, TypeScript, and OTEL/API
  • •identifying gaps between code, tests, and docs

Required outputs

Produce these artifacts in the response (or files if asked):

  1. •Integration inventory table
  2. •Pattern classification per integration
  3. •Cross-surface comparison matrix (Python / TypeScript / OTEL/API)
  4. •Findings list (bugs/risks/gaps)
  5. •Documentation patch plan
  6. •Test plan (unit + integration + e2e)

Workflow

  1. •Build inventory and classify pattern per integration.
  2. •Extract behavior contract (capture, streaming, errors, flush, span hierarchy).
  3. •Extract data contract (usage, metadata/tags, thread propagation).
  4. •Validate backend compatibility (OTEL endpoint/protocol/mapping rules).
  5. •Review tests for coverage and missing edge cases.
  6. •Produce documentation updates with consistent template.

Load references as needed

  • •For matrix schema and scoring: references/matrix-and-rubric.md
  • •For pattern definitions: references/patterns.md
  • •For evidence collection paths and commands: references/evidence-collection.md
  • •For documentation update template: references/doc-template.md
  • •For testing checklist: references/testing-checklist.md
  • •For subagent prompt templates: references/subagent-prompts.md

Inputs

  • •Integration scope (specific integration or full inventory)
  • •Target surfaces to compare (Python, TypeScript, OTEL/API)
  • •Relevant repo paths and branch/diff scope
  • •Output expectations (table depth, doc patch format, test depth)

Outputs

  • •Inventory and comparison matrix
  • •Findings and risk summary
  • •Documentation patch plan
  • •Test plan with missing coverage calls

Default evidence locations (override if repo differs)

  • •Python:
    • •sdks/python/src/opik/integrations
    • •sdks/python/design/INTEGRATIONS.md
    • •sdks/python/design/TESTING.md
  • •TypeScript:
    • •sdks/typescript/design/INTEGRATIONS.md
    • •sdks/typescript/design/TESTING.md
  • •Backend OTEL:
    • •apps/opik-backend/src/main/java/com/comet/opik/api/resources/v1/priv/OpenTelemetryResource.java
    • •apps/opik-backend/src/main/java/com/comet/opik/domain/OpenTelemetryService.java
    • •apps/opik-backend/src/main/java/com/comet/opik/domain/OpenTelemetryMapper.java
    • •apps/opik-backend/src/main/java/com/comet/opik/domain/mapping/OpenTelemetryMappingRuleFactory.java
    • •apps/opik-backend/docs/integrations.md