AI Agent Skills by Confluent is a collection of skills for building Kafka producers, Flink applications, and real-time data-streaming pipelines. Developers use it with coding assistants when creating applications and pipelines on Confluent. The catalogue entries are its skills, plugin, and instruction.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add confluentinc/agent-skills --skill flink-udfgit clone --depth 1 https://github.com/confluentinc/agent-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/confluentinc/agent-skills/flink-udf)<a href="https://agentmods.dev/skills/confluentinc/agent-skills/flink-udf"><img src="https://agentmods.dev/badge/skills/confluentinc/agent-skills/flink-udf/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/confluentinc/agent-skills/flink-udf"><img src="https://agentmods.dev/badge/skills/confluentinc/agent-skills/flink-udf.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00189 | $0.00811 |
| Opus 5 | $0.00095 | $0.00405 |
| Sonnet 5 | $0.00038 | $0.00162 |
| Haiku 4.5 | $0.00019 | $0.00081 |
Grade A, and why
flink-udf scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Flink User-Defined Functions (UDFs)
Build and deploy custom functions in Java for Apache Flink to extend SQL and Table API capabilities with custom logic.
Function Types
Before proceeding, identify which type of function the user needs:
- Scalar UDF: Maps input values to a single output value (e.g., custom hash, string manipulation, calculations)
- User-Defined Table Function (UDTF): Maps input to multiple output rows (e.g., split strings, explode arrays)
- Process Table Function (PTF): Advanced stateful processing with N-to-M semantics, managed state, and timers (e.g., windowing, deduplication, state machines)
Gather Requirements
Ask the user these questions to determine the implementation path (if not already clear from context):
- Deployment target: Confluent Cloud or local Docker?
- Infrastructure: Deploy new infrastructure (Kafka + Flink) or use existing?
- Invocation method: Flink SQL or Table API?
Route to Implementation Guide
Based on the answers above, read the appropriate reference file:
Confluent Cloud Deployment
- Scalar UDF or UDTF → Read
references/udf-udtf-java-confluent-cloud.md - Process Table Function (PTF) → Read
references/ptf-java-confluent-cloud.md
If infrastructure setup is needed, also read references/confluent-cloud-setup.md first.
Local Docker Deployment
- Scalar UDF or UDTF → Read
references/udf-udtf-java-local.md - Process Table Function (PTF) → Read
references/ptf-java-local.md
If infrastructure setup is needed, also read references/local-docker-setup.md first.
Implementation Workflow
After reading the appropriate reference:
- Set up infrastructure (if needed)
- Generate boilerplate code for the function
- Implement the business logic
- Build and package the JAR
- Confirm the deployment plan with the user. Before any resource-modifying call (
confluent flink artifact create,docker cpinto a running container,CREATE FUNCTION, etc.), present the plan and wait for explicit approval. Show:- Artifact name and JAR path
- Function name to register
- Target environment (Confluent Cloud env + compute pool ID, or local Docker container name)
- The exact commands and SQL that will run Do not proceed to steps 6–7 until the user confirms.
- Deploy the artifact
- Register the function in Flink
- Test the function with sample data
- Provide usage examples (SQL or Table API)
What ships with it
9 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- evals/evals.json 11 KB
- references/confluent-cloud-setup.md 6.6 KB
- references/dependencies-confluent-cloud.md 3.3 KB
- references/dependencies-local.md 3.5 KB
- references/local-docker-setup.md 8.3 KB
- references/ptf-java-confluent-cloud.md 11 KB
- references/ptf-java-local.md 9.9 KB
- references/udf-udtf-java-confluent-cloud.md 7.2 KB
- references/udf-udtf-java-local.md 6.6 KB
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 66 lines · 189 tokens per session scan A faa794eca039
flink-udf is a skill published in the GitHub repository confluentinc/agent-skills (56 stars, last pushed 6d ago), licensed Apache-2.0. It adds 189 tokens to every session and 811 once invoked, about $0.0009 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
developing-genkit-js
Develop AI-powered applications using Genkit in Node.js/TypeScript. Use when the user asks about Genkit, AI agents, flows, or tools in JavaScript/TypeScript, or when encountering Genkit errors, validation issues, type errors, or API problems.
developing-genkit-python
Develop AI-powered applications using Genkit in Python. Use when the user asks about Genkit, AI agents, flows, or tools in Python, or when encountering Genkit errors, import issues, or API problems.
azure-ai-vision-imageanalysis-java
Build image analysis applications with Azure AI Vision SDK for Java. Use when implementing image captioning, OCR text extraction, object detection, tagging, or smart cropping.
tsp-type-override
Override TypeSpec types with Java-native types (e.g. OffsetDateTime, DayOfWeek) using @@alternateType in a client.java.tsp file. Use when a TypeSpec model field has an incorrect or too-generic type that should map to a specific Java type.
tsp-naming-collision
Fix Java codegen parameter names that end with a numeric suffix (e.g. createAgentRequest1) caused by TypeSpec model names colliding with synthetic body type names. Use when generated Java client methods have parameter names ending in '1'.
dup-classes
Verify whether generated Java classes duplicate openai-java models by comparing fields/types (names may differ). Use when checking for duplicate model coverage.