skills.rest
Cloud-Byte-Consulting
Cloud-Byte-ConsultingOFFICIAL
Explore all skills by this author →
GitHub

data-contract-author

Draft and validate ODCS data contracts with sensitivity tagging, semver policy, and CI enforcement.

Updated Jul 16, 2026
Search Tags:#semver#schema validation#data contract#odcs#ci enforcement#data contract cli#pii tagging
One-click install
npx skills add https://github.com/Cloud-Byte-Consulting/plugins --skill data-contract-author-cloud-byte-consulting
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: data-contract-author
Source: https://github.com/Cloud-Byte-Consulting/plugins/tree/main/research-data-platform/skills/data-contract-author
Command: npx skills add https://github.com/Cloud-Byte-Consulting/plugins --skill data-contract-author-cloud-byte-consulting

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Upstream dataset changes keep surprising training and evaluation teams because schema agreements live in wiki pages that decay. This Skill turns those informal agreements into versioned, machine-verifiable YAML data contracts that are executed in CI, so breaking changes fail a pipeline instead of silently corrupting a training run. ## Core Features & Use Cases - ODCS Contract Authoring: Drafts section-by-section ODCS v3 contracts covering schema, semantics, quality rules, SLAs, team lineage, and custom properties like snapshot IDs and artifact manifests. - Sensitivity and Semantic Tagging: Applies field-level sensitivity classifications and schema.org semantic tags so policy engines and agents can act on PII and meaning. - CI Enforcement with Data Contract CLI: Wires lint, test, breaking-change diff, and export stages into the dataset repo pipeline, with a semver-based breaking-change policy. - Use Case: A curation team hands a curated instruction-tuning corpus to the training team. Use this Skill to author a contract pinning the schema, label semantics, dedup keys, and snapshot addressing, then gate contract PRs on datacontract breaking so any redefinition of a label class requires a major version bump. ## Quick Start Use the data-contract-author skill to draft an ODCS contract for the curated instruction-tuning dataset handed off from curation to training, including PII tags and a CI snippet.

Frequently Asked Questions about data-contract-author

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write a data contract for a dataset handoff between teams?▼

Author one ODCS YAML contract per dataset and version, covering schema, semantics, quality rules, SLAs, and team ownership. Place contracts only at team or system boundaries, elicit concrete example records first, then enforce the contract in CI with Data Contract CLI.

ODCS vs JSON Schema for data contracts, which should I use?▼

Author in ODCS because it prompts for semantics, SLAs, roles, and quality rules that plain JSON Schema lacks. Operate through Data Contract CLI, which reads ODCS, and generate JSON Schema, pydantic, or dbt artifacts from the contract for consumers.

How do I enforce data contracts in CI with Data Contract CLI?▼

Run datacontract lint on every PR, datacontract test against the actual store post-ingest, datacontract breaking to detect breaking changes between versions, and datacontract export to generate consumer artifacts like JSON Schema and SQL DDL.

What counts as a breaking change in a data contract?▼

Changing a column type or name, removing a column or key, merging tables, or redefining label semantics are breaking changes requiring a major version bump. Adding optional columns or new label classes is minor; metadata edits are patches.

When should I not create a data contract?▼

Skip contracts for internal plumbing within one team and for exploratory scratch data. Contracts pay off only where a second team or an agent consumes the data, since the definition and enforcement effort needs a boundary to justify it.

How do I tag PII and sensitivity in a data contract?▼

Annotate every field, including nested fields, with a sensitivity classification (public, internal, confidential, restricted, not-sensitive) and a schema.org semantic tag. State scrubbing guarantees, such as email redaction rates, as enforceable quality rules rather than informal promises.

Related Skills

Explore More Data & Analytics Skills→
hoavdchoavdc
21

codexkit-contract-drafter

Draft commercial contracts from a clause library with structured sections and risk annotations.

Community
Advanced
SunshinePeace213SunshinePeace213

customer-discovery-synthesis

Synthesize customer discovery interviews and score evidence against kill criteria.

Community
Advanced
Sir-chawakornSir-chawakorn
9

validate-data-quality

Execute rule-based data quality checks for completeness, uniqueness, and referential integrity.

Community
Advanced
anotherbenanotherben
6

full-cycle-fast

Automate Analyze, Plan, Build, Test, and Ship workflows for small code changes.

Community
Intermediate