generated: '2026-08-27'
method: searched
source: https://github.com/confluentinc/agent-skills
description: >-
  Confluent PUBLISHES its own Agent Skills — this is not a set API Evangelist generated. The
  repository is vendored under skills/agent-skills/ verbatim. Confluent states the skills follow
  the Agent Skills Specification at https://agentskills.io/specification, and distributes them
  through a Claude Code plugin marketplace as well as the Skills CLI for Cursor, GitHub Copilot
  and any other spec-compatible agent.
  Because the provider authored these, nothing here is generated on Confluent's behalf.
publisher: Confluent
repo: https://github.com/confluentinc/agent-skills
docs: https://docs.confluent.io/cloud/current/ai/ai-tools/agent-skills.html
specification:
  name: Agent Skills Specification
  url: https://agentskills.io/specification
distribution:
  - kind: claude-code-plugin
    marketplace: confluent-agent-skills
    plugin: streaming-skills-plugin
    file: skills/agent-skills/.claude-plugin/marketplace.json
  - kind: cursor-plugin
    file: skills/agent-skills/.cursor-plugin/plugin.json
  - kind: skills-cli
    note: For Cursor, GitHub Copilot and other spec-compatible agents.
supported_agents: [Claude Code, Cursor, GitHub Copilot]
local_path: skills/agent-skills/
license_file: skills/agent-skills/LICENSE
skill_count: 9
skills:
  - name: confluent-cloud-cdc-tableflow
    path: skills/agent-skills/skills/confluent-cloud-cdc-tableflow/SKILL.md
    summary: >-
      End-to-end Change Data Capture pipelines on Confluent Cloud using Debezium source
      connectors, Flink for transformation, and Tableflow for Iceberg/Delta Lake integration.
    prerequisites: Confluent Cloud account, open-source MCP server or Confluent CLI
    has_evals: true
  - name: confluent-skill-creator
    path: skills/agent-skills/skills/confluent-skill-creator/SKILL.md
    summary: >-
      Build and test new Confluent product skills end to end — scoping, drafting,
      spec-compliance validation, and end-to-end tests against a real Confluent environment.
    prerequisites: Python 3.9+, confluent-kafka with Avro/JSON Schema/Protobuf extras, a Confluent environment
    has_evals: true
  - name: confluent-skill-reviewer
    path: skills/agent-skills/skills/confluent-skill-reviewer/SKILL.md
    summary: >-
      Audit a Confluent agent skill against the Agent Skills Specification, Confluent
      conventions and eval coverage requirements; returns blocking issues, warnings and nits.
    has_evals: true
  - name: developing-kafka-java-client
    summary: >-
      Scaffold or extend Java Kafka producer and consumer projects with Schema Registry
      serialization (Avro, JSON Schema, Protobuf) for Confluent Cloud, local Docker or
      WarpStream, using Maven or Gradle.
    prerequisites: Java 17+, Maven or Gradle, Docker for local development
    note: >-
      Listed in Confluent's published skill table on docs.confluent.io. Not present in the
      vendored snapshot of the repository, which was cloned before this skill landed.
  - name: developing-kafka-python-client
    path: skills/agent-skills/skills/developing-kafka-python-client/SKILL.md
    summary: >-
      Scaffold Python Kafka producer and consumer projects using confluent-kafka-python with
      Schema Registry serialization, in asynchronous or synchronous mode.
    prerequisites: Python, pip, Docker for local development
    has_evals: true
  - name: flink-udf
    path: skills/agent-skills/skills/flink-udf/SKILL.md
    summary: >-
      Build and deploy Flink user-defined functions in Java — scalar UDFs, UDTFs and process
      table functions — invoked from Flink SQL or the Table API on Confluent Cloud or Docker.
    prerequisites: Java 17+, Maven or Gradle, Confluent CLI or Docker
    has_evals: true
  - name: kafka-schema-registry
    path: skills/agent-skills/skills/kafka-schema-registry/SKILL.md
    summary: >-
      Scan projects for Kafka applications, extract schemas from data models, tag PII fields,
      generate Terraform for Schema Registry registration, and produce a migration report with
      rollout ordering.
    prerequisites: Terraform
    has_evals: true
  - name: kafka-streams-programming
    path: skills/agent-skills/skills/kafka-streams-programming/SKILL.md
    summary: >-
      Architect, build and debug Kafka Streams applications — topology design, pattern
      selection (joins, windows, aggregations), code generation and production troubleshooting.
    prerequisites: Java 17+, Gradle or Maven, Docker
    has_evals: true
  - name: msk-migration
    summary: >-
      Assess and plan a migration from Amazon MSK to Confluent Cloud, including an environment
      readiness audit and a technical migration plan covering sizing, networking,
      authentication and switchover.
    prerequisites: An AWS MSK environment; KCP (optional) for automated scans
    note: >-
      Listed in Confluent's published skill table on docs.confluent.io. Not present in the
      vendored snapshot of the repository.
quality_signals:
  evals_as_contract: >-
    Every vendored skill ships an evals/evals.json alongside SKILL.md, and the repository
    enforces it — confluent-skill-reviewer audits eval coverage as one of three named gates.
    This is a materially stronger publishing posture than a skills directory alone.
  anti_triggers: >-
    Most skills carry explicit "Do NOT trigger for..." clauses in their description, which is a
    deliberate defence against trigger overlap; the repository ships
    scripts/check_trigger_overlap.py to test it.
  lazy_loading: >-
    Skills keep detail in references/ and instruct the agent to load only what it needs, rather
    than inlining everything into SKILL.md.