480-data-engineering

A set of standards for building batch and streaming data pipelines, which process data in groups or continuously. It covers data contracts, quality checks, privacy, monitoring, replaying, and cost control across different data platforms.

In plain words
What is it for?
It helps design repeatable jobs, enforce allowed schema changes, validate data near ingestion and publication, protect personal data and secrets, monitor pipeline runs, and process only the data needed.
Why use it?
It reduces duplicate outputs, unreproducible backfills, unnoticed data errors, privacy failures, hard-to-debug runs, and unnecessary full scans.

Cursor rule

Install

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.

agentmods
npx agentmods add rules/d-padmanabhan/agent-engineering-handbook/480-data-engineering
Clone the repo
git clone --depth 1 https://github.com/d-padmanabhan/agent-engineering-handbook
Per session 26 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,462 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00026 $0.01462
Opus 5 $0.00013 $0.00731
Sonnet 5 $0.00005 $0.00292
Haiku 4.5 $0.00003 $0.00146

Measured 3d ago against content hash 3d9a9124171e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

480-data-engineering 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 3d 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.

rules/480-data-engineering.mdc · 191 lines

How it starts

The opening of the file, as written. The whole thing — 191 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Data Engineering Ruleset

Audience: data engineers and platform engineers building batch + streaming pipelines

Goal: make data systems safe-by-default, reproducible, observable, and cost-aware across engines (Databricks, Snowflake, Kafka, Teradata)

[!NOTE] This rule is cross-platform. Pair it with engine-specific rules:

  • 481-databricks.mdc
  • 482-snowflake.mdc
  • 483-kafka.mdc
  • 484-teradata.mdc

Non-negotiables (defaults)

  • Idempotency: re-running the same job for the same input window must not duplicate or corrupt outputs.
  • Deterministic replays: backfills and replays must be reproducible from inputs + code + config.
  • Schema contracts: define and enforce what changes are allowed (additive vs breaking).
  • Data quality: validate critical invariants close to ingestion and before publishing.
  • Governance: treat PII and secrets as production incidents; enforce least privilege.
  • Observability: every pipeline run emits enough signals to debug quickly.
  • Cost discipline: design for pruning/incremental processing; avoid full scans by default.

Data contracts and schema evolution

Contract basics

  • Publish a stable contract per dataset/topic (schema + semantics + SLA).
  • Include:
    • keys (natural/business keys and/or surrogate keys)
    • event time vs processing time meaning
    • nullability and defaults
    • allowed late-arrival window
    • dedupe strategy (if any)

Schema evolution rules of thumb

  • Prefer additive changes (new nullable column) over destructive changes.
  • Breaking changes require an explicit migration plan:
    • dual-write/dual-read window
    • versioned outputs (v1, v2) or compatibility mode
    • clear rollback strategy

[!WARNING] Changing meaning without changing schema is still a breaking change (for example: units, currency, time zone, enums).


Idempotency patterns (batch + streaming)

“Prove then publish” pattern

  • Write to a staging location/table.
  • Validate row counts and invariants.
  • Publish via an atomic swap/merge (engine-specific).

Read the full file on GitHub · 191 lines

Changes

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.

  1. 3d ago First seen · 191 lines · 1,462 tokens per session scan A 3d9a9124171e

Subscribe to this mod's changes

480-data-engineering is a cursor rule published in the GitHub repository d-padmanabhan/agent-engineering-handbook (15 stars, last pushed 3d ago), licensed MIT. It adds 26 tokens to every session and 1,462 once invoked, about $0.0001 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.