Agent skill

Rocky Poc

Authoring a new POC under examples/playground/pocs/. An agent skill from rocky-data/rocky.

Apache-2.0Auto-check passedData & Analytics

Install Rocky Poc

skills CLI

$ npx skills add rocky-data/rocky --skill rocky-poc -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI

$ gh skill install rocky-data/rocky rocky-poc --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy

$ git clone --depth 1 https://github.com/rocky-data/rocky.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/rocky-poc .claude/skills/rocky-poc && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Paste into Claude Code

Install the "rocky-poc" agent skill from https://github.com/rocky-data/rocky/tree/main/.agents/skills/rocky-poc into .claude/skills/rocky-poc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "rocky-poc", then confirm the skill loads.

Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

$skill-installer

$skill-installer install https://github.com/rocky-data/rocky/tree/main/.agents/skills/rocky-poc

Type this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.

skills CLI

$ npx skills add rocky-data/rocky --skill rocky-poc -a codex

Project install goes to .agents/skills/; add -g for ~/.codex/skills/.

GitHub CLI

$ gh skill install rocky-data/rocky rocky-poc --agent codex

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/rocky-data/rocky.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.agents/skills/rocky-poc .agents/skills/rocky-poc && rm -rf skills-src

Use ~/.agents/skills/ instead of .agents/skills for a personal install.

Codex skills documentation · loads skills from .agents/skills/

Paste into Codex

Install the "rocky-poc" agent skill from https://github.com/rocky-data/rocky/tree/main/.agents/skills/rocky-poc into .agents/skills/rocky-poc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "rocky-poc", then confirm the skill loads.

Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add rocky-data/rocky --skill rocky-poc -a cursor

Project install goes to .agents/skills/; add -g for ~/.cursor/skills/.

GitHub CLI

$ gh skill install rocky-data/rocky rocky-poc --agent cursor

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/rocky-data/rocky.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.agents/skills/rocky-poc .cursor/skills/rocky-poc && rm -rf skills-src

Use ~/.cursor/skills/ instead of .cursor/skills for a personal install.

Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/

Paste into Cursor

Install the "rocky-poc" agent skill from https://github.com/rocky-data/rocky/tree/main/.agents/skills/rocky-poc into .cursor/skills/rocky-poc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "rocky-poc", then confirm the skill loads.

Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

gemini skills install

$ gemini skills install https://github.com/rocky-data/rocky.git --path .agents/skills/rocky-poc

--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.

skills CLI

$ npx skills add rocky-data/rocky --skill rocky-poc -a gemini-cli

Project install goes to .agents/skills/; add -g for ~/.gemini/skills/.

GitHub CLI

$ gh skill install rocky-data/rocky rocky-poc --agent gemini-cli

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/rocky-data/rocky.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.agents/skills/rocky-poc .gemini/skills/rocky-poc && rm -rf skills-src

Use ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.

Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/

Paste into Gemini CLI

Install the "rocky-poc" agent skill from https://github.com/rocky-data/rocky/tree/main/.agents/skills/rocky-poc into .gemini/skills/rocky-poc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "rocky-poc", then confirm the skill loads.

Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

GitHub CLI

$ gh skill install rocky-data/rocky rocky-poc

Installs for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).

skills CLI

$ npx skills add rocky-data/rocky --skill rocky-poc -a github-copilot

Project install goes to .agents/skills/; add -g for ~/.copilot/skills/.

Manual copy

$ git clone --depth 1 https://github.com/rocky-data/rocky.git skills-src && mkdir -p .github/skills && cp -r skills-src/.agents/skills/rocky-poc .github/skills/rocky-poc && rm -rf skills-src

Use ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.

GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/

Paste into GitHub Copilot

Install the "rocky-poc" agent skill from https://github.com/rocky-data/rocky/tree/main/.agents/skills/rocky-poc into .github/skills/rocky-poc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "rocky-poc", then confirm the skill loads.

GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add rocky-data/rocky --skill rocky-poc -a opencode

OpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.

GitHub CLI

$ gh skill install rocky-data/rocky rocky-poc --agent opencode

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/rocky-data/rocky.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.agents/skills/rocky-poc .opencode/skills/rocky-poc && rm -rf skills-src

Use ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.

OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/

Paste into OpenCode

Install the "rocky-poc" agent skill from https://github.com/rocky-data/rocky/tree/main/.agents/skills/rocky-poc into .opencode/skills/rocky-poc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "rocky-poc", then confirm the skill loads.

OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

Facts

Skill name: rocky-poc
GitHub stars: 304
Token cost: ~1.7k tokens
SKILL.md length: 571 words
Files: 1
Skills in repo: 22
Repo updated: Oct 9, 2026
First seen: Oct 7, 2026
Licence: Apache-2.0

At a glance

Authoring a new POC under examples/playground/pocs/. An agent skill from rocky-data/rocky.

Works in 5 steps: Feature — one-sentence description of… → Why it's distinctive — why a user should… → Layout — tree of files with one-line… → …
Demoing a single Rocky feature (drift
SKILL.md covers When to use this skill, What's already covered in…, POC categories and Scaffold, plus 9 more sections
Calls duckdb; needs DATABRICKS_TOKEN

What it does

Rocky Poc is an agent skill from rocky-data/rocky. Authoring a new POC under examples/playground/pocs/. Use when demoing a single Rocky feature (drift, incremental, contracts, lineage, AI, hooks, etc.) that doesn't already exist in engine/examples/. Enforces the POC conventions so it runs via a single ./run.sh with no credentials.

Its SKILL.md is about 1.7k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Data & Analytics, covering Data pipelines and ETL. It works with DuckDB. The repository describes itself as: A SQL transformation engine that type-checks your whole pipeline and catches breaking changes before they run — branches, replay, column-level lineage, compile-time contracts… The licence is Apache-2.0.

When your agent uses it

Demoing a single Rocky feature (drift
Etc.) that doesnt already exist in engine/examples/

Example prompts

“/rocky-poc”

Requirements

A credential in DATABRICKS_TOKEN

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

Feature — one-sentence description of what this POC demonstrates
Why it's distinctive — why a user should care (what it proves about Rocky vs alternatives)
Layout — tree of files with one-line descriptions
Run — copy-pasteable commands to run it
Expected output — what the user should see (snippet of run.sh output, not the full golden file)

What it can do on your machine

Read from SKILL.md and the folder at commit 9c3d777. It shows what the files ask for, not the result of running them.

Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Shell commands in SKILL.md call:
- duckdb
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names these keys or tokens, usually read from environment variables:
- DATABRICKS_TOKEN
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Rocky Poc loads about 1.7k tokens when it runs. Until then it costs about 74 tokens; SKILL.md has 571 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it: ~74
When it runs · the whole SKILL.md, loaded when a task matches: ~1.7k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from rocky-data/rocky at commit 9c3d777, republished under its Apache-2.0 licence (© rocky-data). 571 words, ~1,721 tokens.

Download SKILL.mdSave it as .claude/skills/rocky-poc/SKILL.md (or your agent's skills folder).

name: rocky-poc
description: Authoring a new POC under `examples/playground/pocs/`. Use when demoing a single Rocky feature (drift, incremental, contracts, lineage, AI, hooks, etc.) that doesn't already exist in `engine/examples/`. Enforces the POC conventions so it runs via a single `./run.sh` with no credentials.

Authoring a new Rocky POC

examples/playground/ is a curated catalog of small, self-contained POCs — one per feature. Each POC is runnable with one ./run.sh, defaults to DuckDB (no credentials), and lives in a category folder.

When to use this skill

Demoing a Rocky feature not already covered in engine/examples/ (the official starter set lives there and should NOT be duplicated in playground)
Building a runnable smoke test for a new capability
Creating a benchmark fixture (anything under benchmarks/ is off-limits unless explicitly asked)

What's already covered in `engine/examples/`

Do NOT build a duplicate playground POC for these — link to them instead:

`engine/examples/` path	What it shows
`quickstart/`	3-model "hello world" pipeline
`multi-layer/`	Generic Bronze/Silver/Gold
`dbt-migration/`	dbt → Rocky before/after
`ai-intent/`	AI-intent feature walkthrough
`dagster-integration/`	Basic Dagster wiring

Playground POCs should cover what those don't: incremental watermarks, drift diagnostics, contract validation, hooks, lineage, AI sync, custom adapters, partitions, column lineage, etc.

POC categories

Pick the category whose id matches the feature group:

00-foundations          Fundamental concepts, playground baseline
01-quality              Contracts, checks, drift, data quality
02-performance          Incremental, partitioning, checksums, adaptive concurrency
03-ai                   AI intent, AI sync, AI test, AI explain
04-governance           Permissions, tags, workspace isolation
05-orchestration        Dagster, sensors, schedules, hooks
06-developer-experience LSP, lineage, VS Code features
07-adapters             Adapter-specific behavior

Scaffold

Use the scaffolder — don't hand-create the directory:

bash

cd examples/playground
./scripts/new-poc.sh <category> <id-name>
# e.g.
./scripts/new-poc.sh 02-performance 07-late-arriving-data

This copies scripts/_poc-template/ into pocs/<category>/<id-name>/ and chmods run.sh.

POC structure

Every POC MUST contain:

README.md         # Structured: feature, why distinctive, layout, run, expected output
rocky.toml        # Minimal POC-specific config (DuckDB by default)
run.sh            # Executable end-to-end demo (chmod +x)
models/           # .sql / .rocky files + .toml sidecars
contracts/        # Only if contracts are part of the feature
seeds/            # CSV / SQL sample data — keep ≤1000 rows
data/seed.sql     # Optional — auto-loaded by `rocky test`
expected/         # Golden JSON output from run.sh (gitignored, regenerated each run)

Minimal `rocky.toml` (DuckDB)

toml

[adapter]
type = "duckdb"
path = "poc.duckdb"

[pipeline.poc]
strategy = "full_refresh"   # or "incremental"

[pipeline.poc.source]
schema_pattern = { prefix = "raw__", separator = "__", components = ["source"] }

[pipeline.poc.target]
catalog_template = "poc"
schema_template = "staging__{source}"

Defaults that can be omitted:

pipeline.type defaults to "replication" — omit it.
Unnamed [adapter] with a type key auto-wraps as adapter.default; pipeline adapter refs default to "default" — omit adapter = "local" lines.
[state]\nbackend = "local" is the default — omit it.
auto_create_catalogs = false / auto_create_schemas = false are defaults — omit unless intentionally true.
Model sidecar name defaults to filename stem; target.table defaults to name — omit when redundant.
models/_defaults.toml provides directory-level [target] defaults (catalog, schema) — use when 2+ models share values.

Credential gating

Default to DuckDB so the POC runs with zero config. Only require credentials when the feature genuinely can't be demoed without them (Databricks governance, Snowflake dynamic tables, Fivetran, Anthropic API).

If credentials are required, fail fast at the top of run.sh:

bash

: "${DATABRICKS_HOST:?Set DATABRICKS_HOST before running}"
: "${DATABRICKS_TOKEN:?Set DATABRICKS_TOKEN before running}"

And mark the credential requirement prominently in README.md.

Show full SKILL.md (261 more words)Show less

Runtime idioms

Goal	Command
Type-check models without a warehouse	`rocky compile --models models/ --contracts contracts/`
Run model tests against in-memory DuckDB (auto-loads `data/seed.sql`)	`rocky test --models models/ --contracts contracts/`
CI pipeline (compile + test)	`rocky ci --models models/ --contracts contracts/`
Validate the pipeline config	`rocky validate -c rocky.toml`
Discover sources from local DuckDB	`rocky -c rocky.toml discover`
Preview replication SQL	`rocky -c rocky.toml plan --filter source=<name>`
Execute the pipeline	`rocky -c rocky.toml run --filter source=<name>`
Inspect watermarks	`rocky -c rocky.toml state`
Column-level lineage	`rocky lineage <model> --models models/ [--column <col>]`
Schema-pattern aware diagnostics	`rocky doctor`

For discover/plan/run, seed data must be manually loaded into the DuckDB file first — typically duckdb poc.duckdb < data/seed.sql at the top of run.sh. rocky test auto-loads it.

`run.sh` conventions

set -euo pipefail
Print a header describing the feature
Run the canonical command sequence (compile → test → run → inspect)
Exit 0 on success; the POC is smoke-tested by ./scripts/run-all-duckdb.sh with a 60s timeout

Verification before committing

bash

cd pocs/<cat>/<id-name>
./run.sh                          # Must exit 0
rocky validate -c rocky.toml      # If the POC uses a pipeline path

For the catalog as a whole:

bash

cd examples/playground
./scripts/run-all-duckdb.sh       # All credential-free POCs, 60s timeout each

README structure

Five sections, in this order:

Feature — one-sentence description of what this POC demonstrates
Why it's distinctive — why a user should care (what it proves about Rocky vs alternatives)
Layout — tree of files with one-line descriptions
Run — copy-pasteable commands to run it
Expected output — what the user should see (snippet of run.sh output, not the full golden file)

Commit style

feat(02-performance/07-late-arriving-data): add late-arriving watermark POC

Scope by POC id when the change is POC-specific.

Off-limits

benchmarks/ — Rocky vs dbt-core / dbt-fusion / PySpark perf suite. Don't touch unless explicitly asked.
Duplicating anything already in engine/examples/. Link to it instead.

Files

Just SKILL.md in .agents/skills/rocky-poc of rocky-data/rocky.

Open the folder on GitHubat commit 9c3d777

Compare with similar skills

Rocky Poc next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Rocky Poc compared with similar skills
Skill	Stars	Used in	Tokens	Auto-check	Licence	Repo updated
Rocky Poc this skillrocky-data/rocky	304	—	~1.7k	Automated safety check: Pass	Apache-2.0	yesterday
Dinobase Business Data Querieskappa90/dinobase	263	—	~1.5k	Automated safety check: Pass	Custom licence	3 mo ago
Windmill Data Pipeline Authorwindmill-labs/windmill	18k	—	~4k	Automated safety check: Pass	Custom licence	today
Erd Studio Setupliam-machine/erd-studio	165	—	~8.5k	Automated safety check: Pass	Custom licence	yesterday
Tushare Plugin BuilderYourdaylight/stock_datasource	189	—	~2.5k	Automated safety check: Pass	MIT	1 mo ago
Centia Snapshot Catalogmapcentia/geocloud2	152	—	~2.3k	Automated safety check: Pass	AGPL-3.0	yesterday

Similar skills

Dinobase Business Data Queries
kappa90/dinobase
Sets up Dinobase, a local DuckDB database that syncs data from 100+ business sources, then answers questions across them with SQL joins and previewed write-backs.
263 GitHub stars~1.5k tokensUpdated 3 mo ago
Data & AnalyticsAuto-check passed
Windmill Data Pipeline Author
windmill-labs/windmill
Builds Windmill data pipelines as independent, pipeline-annotated scripts forming a DAG over shared storage, defaulting to DuckDB nodes that materialize into DuckLake tables.
18k GitHub stars~4k tokensUpdated today
Data & AnalyticsAuto-check passed
Erd Studio Setup
liam-machine/erd-studio
Friendly, step-by-step setup for ERD Studio in an existing dbt project, for people who may be new to dbt or data modelling.
165 GitHub stars~8.5k tokensUpdated yesterday
Data & AnalyticsAuto-check passed
Tushare Plugin Builder
Yourdaylight/stock_datasource
Turns a Tushare API doc URL into a full data plugin for the stock_datasource repo: extractor, ClickHouse schema, query service, config and curl examples.
189 GitHub stars~2.5k tokensUpdated 1 mo ago
Data & AnalyticsAuto-check passed
Centia Snapshot Catalog
mapcentia/geocloud2
Analyse GC2/Centia Parquet snapshots with DuckDB by walking the STAC catalog.json in the snapshot store — find datasets, decide whether a dataset has geometry (and in which CRS), read one snapshot…
152 GitHub stars~2.3k tokensUpdated yesterday
Data & AnalyticsAuto-check passed
Dinobase Connector Builder
kappa90/dinobase
Writes a new Dinobase YAML connector for a REST API that has no verified dlt source, covering auth, pagination, read and write endpoints and incremental loading.
263 GitHub stars~1.9k tokensUpdated 3 mo ago
Backend & APIsAuto-check passed

Questions about Rocky Poc

What does Rocky Poc do?

Authoring a new POC under examples/playground/pocs/. An agent skill from rocky-data/rocky. Rocky Poc is an agent skill from rocky-data/rocky. Authoring a new POC under examples/playground/pocs/.

When should I use Rocky Poc?

Rocky Poc fits situations like: demoing a single Rocky feature (drift; etc.) that doesnt already exist in engine/examples/.

How do I install Rocky Poc in Claude Code?

Run `npx skills add rocky-data/rocky --skill rocky-poc -a claude-code`. Or copy the skill folder (.agents/skills/rocky-poc in rocky-data/rocky) into .claude/skills/rocky-poc in your project. Claude Code loads it when a task matches its description.

How do I install Rocky Poc in Codex?

Run `npx skills add rocky-data/rocky --skill rocky-poc -a codex`. Or copy the skill folder (.agents/skills/rocky-poc in rocky-data/rocky) into .agents/skills/rocky-poc in your project. Codex loads it when a task matches its description.

Can I use Rocky Poc in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add rocky-data/rocky --skill rocky-poc -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/rocky-poc, .gemini/skills/rocky-poc, .github/skills/rocky-poc and .opencode/skills/rocky-poc in your project.

What does Rocky Poc need to run?

Going by SKILL.md and its folder, Rocky Poc needs the command-line tools its instructions call (duckdb) and credentials named DATABRICKS_TOKEN. Our summary lists: A credential in DATABRICKS_TOKEN.

Does Rocky Poc access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Rocky Poc safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Rocky Poc use?

Rocky Poc is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Rocky Poc use?

About 1.7k tokens (SKILL.md is roughly 6.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Rocky Poc?

Skills that share tags, products or a category with Rocky Poc: Dinobase Business Data Queries (kappa90/dinobase, 263 stars), Windmill Data Pipeline Author (windmill-labs/windmill, 18k stars), Erd Studio Setup (liam-machine/erd-studio, 165 stars) and Tushare Plugin Builder (Yourdaylight/stock_datasource, 189 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Rocky Poc?

rocky-data (a GitHub organization) maintains it in rocky-data/rocky, which has 304 GitHub stars. The repository holds 22 skills in this directory. The repository was last updated on October 9, 2026.

Source: rocky-data/rocky on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.

Install Rocky Poc

Facts

At a glance

What it does

When your agent uses it

Example prompts

Requirements

Workflow steps

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Context cost

Safety

SKILL.md

Authoring a new Rocky POC

When to use this skill

What's already covered in engine/examples/

POC categories

Scaffold

POC structure

Minimal rocky.toml (DuckDB)

Credential gating

Runtime idioms

run.sh conventions

Verification before committing

README structure

Commit style

Off-limits

Files

Compare with similar skills

Similar skills

More from rocky-data/rocky

Related

Questions about Rocky Poc

What's already covered in `engine/examples/`

Minimal `rocky.toml` (DuckDB)

`run.sh` conventions