What is ade-bench-cross-db-tasks?

Guidance for creating or debugging ADE-Bench tasks that work with DuckDB and Snowflake, two database systems with different SQL behavior.

Use it when writing migrations, task setup scripts, database-specific patches, or solution patches for cross-database benchmarks.

It prevents setup and solution patches from being applied in the wrong order or against the wrong database state.

ade-bench-cross-db-tasks — Skill for Claude Code and Codex by dbt-labs

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.

Any agent

npx skills add dbt-labs/ade-bench --skill ade-bench-cross-db-tasks

Clone the repo

git clone --depth 1 https://github.com/dbt-labs/ade-bench

Made for: Claude Code, Codex.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/dbt-labs/ade-bench/ade-bench-cross-db-tasks/github.svg)](https://agentmods.dev/skills/dbt-labs/ade-bench/ade-bench-cross-db-tasks)

Your own site

<a href="https://agentmods.dev/skills/dbt-labs/ade-bench/ade-bench-cross-db-tasks"><img src="https://agentmods.dev/badge/skills/dbt-labs/ade-bench/ade-bench-cross-db-tasks/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.

Your own site · 80×15

<a href="https://agentmods.dev/skills/dbt-labs/ade-bench/ade-bench-cross-db-tasks"><img src="https://agentmods.dev/badge/skills/dbt-labs/ade-bench/ade-bench-cross-db-tasks.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 40 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,985 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits

NVIDIA SkillSpector pass 7 Sept 2026

How audits are shown

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.

Model	Per session	Once invoked
Fable 5.1	$0.00040	$0.01985
Opus 5.5	$0.00016	$0.00794
Sonnet 5.5	$0.00008	$0.00397
Haiku 4.5	$0.00004	$0.00198

Measured 2mo ago against content hash 2d9a3420e937, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

ade-bench-cross-db-tasks 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 2mo 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.

skills/ade-bench-cross-db-tasks/SKILL.md · 160 lines

How it starts

The opening of the file, as written. The whole thing — 160 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ADE-Bench Cross-DB Task Authoring

Overview

Tasks in ade-bench can target DuckDB (default) or Snowflake. A task's files go through a deterministic four-stage lifecycle. Patches that ignore this order produce context mismatches or "reversed patch" prompts that block execution.

Container Lifecycle (in order)

1. MIGRATION   shared/migrations/<name>/migration.sh + migration.patch
               — Converts shared project files from DuckDB to Snowflake syntax
2. SETUP       tasks/<id>/setup.sh  +  tasks/<id>/setup/changes.patch
                                       tasks/<id>/setup/changes.snowflake.patch (optional)
               — Creates the puzzle state the agent must solve
3. AGENT       (agent works here)
4. SOLUTION    tasks/<id>/solution.sh  +  tasks/<id>/solutions/changes.patch
                                          tasks/<id>/solutions/changes.snowflake.patch (optional)

Critical: Migration runs before setup and solution. Any line migration touches is in its post-migration state when patches run.

Patch File Pattern

Always apply the base patch, then conditionally apply the snowflake delta. This applies to both setup.sh and solution.sh:

patch -p1 < /app/setup/changes.patch           # always (use /sage/solutions/ for solution.sh)

if [[ "$*" == *"--db-type=snowflake"* ]]; then
    patch -p1 < /app/setup/changes.snowflake.patch  # additive delta only
fi

Never use if/else to choose between two alternative patches. The base patch handles DuckDB; the snowflake patch stacks on top for Snowflake.

changes.snowflake.patch is a delta — it only contains changes that differ from the DuckDB result. It never replicates what migration already did.

Migration Guidelines

Migration's job: Convert DuckDB-specific syntax to Snowflake equivalents across all tasks sharing a project.

Schema names: schema: main → schema: public
Date functions: STRPTIME(...) → TO_DATE(TO_TIMESTAMP(...))
Generator functions, STRFTIME, etc.

Read the full file on GitHub · 160 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.

2mo ago First seen · 160 lines · 40 tokens per session scan A 2d9a3420e937

Subscribe to this mod's changes

ade-bench-cross-db-tasks is a skill published in the GitHub repository dbt-labs/ade-bench (125 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 40 tokens to every session and 1,985 once invoked, about $0.0002 per session on Opus 5.5. 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.

Databases Debugging

ade-bench-cross-db-tasks next to 6 skills from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
ade-bench-cross-db-tasksdbt-labs/ade-bench	Skill	125	1mo ago	Apache-2.0	A	40	Claude Code, Codex
reproduce-buggoogleapis/mcp-toolbox	Skill	17k	yesterday	Apache-2.0	A	130	Claude Code, Codex
triage-issuesgoogleapis/mcp-toolbox	Skill	17k	yesterday	Apache-2.0	A	164	Claude Code, Codex
postgresql-indexingprowler-cloud/prowler	Skill	15k	yesterday	Apache-2.0	A	108	Claude Code
graphjin-evaldosco/graphjin	Skill	3.2k	5d ago	Apache-2.0	A	27	Claude Code, Codex
axiom-audit-grdb-performanceCharlesWiltgen/Axiom	Skill	1.3k	yesterday	MIT	A	43	Codex
django-perf-reviewgetsentry/skills	Skill	1.1k	yesterday	Apache-2.0	A	55	Claude Code