Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/Galius5136/databricks-spark-3.5-cert-prepnpx agentmods add skills/galius5136/databricks-spark-3.5-cert-prep/spark-connectWrote 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/galius5136/databricks-spark-3.5-cert-prep/spark-connect)<a href="https://agentmods.dev/skills/galius5136/databricks-spark-3.5-cert-prep/spark-connect"><img src="https://agentmods.dev/badge/skills/galius5136/databricks-spark-3.5-cert-prep/spark-connect/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/galius5136/databricks-spark-3.5-cert-prep/spark-connect"><img src="https://agentmods.dev/badge/skills/galius5136/databricks-spark-3.5-cert-prep/spark-connect.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00080 | $0.02772 |
| Opus 5 | $0.00040 | $0.01386 |
| Sonnet 5 | $0.00016 | $0.00554 |
| Haiku 4.5 | $0.00008 | $0.00277 |
Grade A, and why
spark-connect scanned grade A with 1 finding 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
> To regenerate local snapshots: `curl -sSL <url> -o source.html` (or use WebFetch with extraction prompts) for each URL above. How it starts
The opening of the file, as written. The whole thing — 203 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Spark Connect — Exam-Prep Knowledge Base
Source Spark version: 3.5.7 (latest 3.5.x docs) + 3.4 launch context | Chapters: 5 | Generated: 2026-05-24
Scope rule: Anything from Spark 3.4 ≤ version ≤ 3.5.x is in scope for the Databricks Certified Associate Developer for Apache Spark exam (Sec 6). Anything from Spark 4.x is flagged ⚠️ as out of exam scope. The skill is conservative — when 4.x docs were used, content is marked accordingly.
How to Use This Skill
- Without arguments — loads the Core Frameworks below (Spark Connect at exam depth).
- By topic — ask about
gRPC,unresolved logical plan,sc:// URL,SPARK_REMOTE,start-connect-server.sh, etc. → I find and read the relevant chapter. - By chapter —
ch01,ch02,ch03,ch04,ch05.
Core Frameworks & Mental Models
Spark Connect in one paragraph
Spark Connect = decoupled client-server architecture for Apache Spark, introduced in Spark 3.4 and matured for Scala in 3.5. The client (any language) builds a DataFrame query, serializes it as an unresolved logical plan in protocol buffers, ships it over gRPC to a Spark Connect server embedded in the Spark driver. The server resolves the plan against its Catalog, runs Catalyst + Tungsten as normal, and streams results back as Apache Arrow row batches. Result: thin clients, isolated client failures, independent upgrades.
Why it exists (4 pain points it solves)
- Stability — a misbehaving user app can't crash the shared driver
- Upgradability — driver upgrades independently of clients (stable protobuf protocol)
- Debuggability — client is a normal local process you can attach a debugger to
- Language flexibility — protocol is language-agnostic; multiple language clients possible
The wire protocol (memorize for the exam)
- Unresolved logical plans as the language-agnostic protocol
- Encoded with Protocol Buffers
- Transported over gRPC (HTTP/2)
- Results return as Apache Arrow row batches
What ships with it
8 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.
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 · 203 lines · 80 tokens per session scan A 532355291556
spark-connect is a skill published in the GitHub repository Galius5136/databricks-spark-3.5-cert-prep (12 stars, last pushed 3mo ago), licensed MIT. It adds 80 tokens to every session and 2,772 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
alt-import
Parse an Exam Radar (OPTIMETA Alt plugin) export and fold its lecture-emphasis exam-probability signal into the PAIDEIA course index — write course-index/radar.md, annotate course-index/coverage.md with a lecture-emphasis column and divergence flags, and seed a gold-zone weakmap. Invoked by /paideia:alt. The export…
fs-notebook-tabs
A computer-science capstone: an on-device ML keyboard that predicts next words privately — problem, method, evaluation, and defense answers. Built as a decision-grade coursework defense deck for professor, defense committee.
implementing-llms-litgpt
Implements and trains LLMs using Lightning AI's LitGPT with 20+ pretrained architectures (Llama, Gemma, Phi, Qwen, Mistral). Use when need clean model implementations, educational understanding of architectures, or production fine-tuning with LoRA/QLoRA. Single-file implementations, no abstraction layers.
rwkv-architecture
RNN+Transformer hybrid with O(n) inference. Linear time, infinite context, no KV cache. Train like GPT (parallel), infer like RNN (sequential). Linux Foundation AI project. Production at Windows, Office, NeMo. RWKV-7 (March 2025). Models up to 14B parameters.
setup
Configure MLflow tracing for Claude Code.
learning-notes-automation
A workflow for turning videos, podcasts, and articles into structured learning notes. It extracts key ideas and creates flashcards that can be imported into Anki, a spaced-repetition study app.