DBs Framework — AI skill for Claude Code

DBS Framework

**Direction / Blueprints / Solutions** — a platform-neutral architecture for reusable AI skills, agents, and workflow systems.

DBS separates workflow decisions, specialized knowledge, and execution. Maintain one core and adapt it to **Claude, ChatGPT, and Codex** through small platform adapters. This is an authoring framework, not a hosted agent runtime or scheduler.

**Origin:** this is an unofficial adaptation of [The DBS Framework Skill by AI Foundations](https://aifoundations.io/resources/the-dbs-framework-skill). Credit for the original DBS framework and source skill belongs to AI Foundations. Developed by this repository's maintainer from their original work, this version adds platform-neutral guidance, separate adapters, modes, and tooling; it is not an official AI Foundations release or an endorsed project.

**Attribution and permission:** the maintainer confirms that AI Foundations permits publication of this adaptation with a link to their website and clear source credit. See [ATTRIBUTION.md](ATTRIBUTION.md) for that record and [LICENSE](LICENSE) for the separate MIT scope covering new tooling.

Start here

1. Read SKILL.md or provide it to your AI environment.
2. Use the Claude, ChatGPT, or Codex adapter for setup and capability mapping.
3. Ask: “Use DBS in Fast Mode to build a reusable monthly report workflow from a supplied CSV. Validate totals and deliver a local PDF. Do not publish it.”
4. Review the resulting contract and run representative tests before relying on it.

For a plain chat, explicitly provide the instructions and required references. For a skill-aware host, install the package using that host's supported location or import flow. **Adapters are guidance, not installed integrations.** Tools and permissions depend on the actual environment. See [validation status](docs/VALIDATION