DBS Framework: open-source reusable AI skills across Claude, ChatGPT, and Codex
Yasirres · reddit · 2026-10-05
A developer open-sourced DBS Framework, tackling workflow lock-in to a single AI platform by separating capabilities into three platform-neutral layers:
- Direction — when/what to run, workflow, constraints, acceptance criteria.
- Blueprints — domain knowledge: schemas, business rules, style guides, examples, references.
- Solutions — execution layer: native tools, MCP servers, APIs, scripts, browser/computer use.
- Small adapters target Claude, ChatGPT, and Codex instead of duplicating business logic per platform.
- Progressive disclosure keeps context lean: the core skill points to reference files only when needed.
- Repo includes a neutral SKILL.md, three adapters, templates, scaffolding, structural validation, automated tests, Fast/Interactive modes, and guidance on permissions, retries, and observable acceptance criteria.
- Explicitly not an agent runtime/scheduler — it's an authoring/architecture framework executing via the host's own tools.
Unofficial adaptation of the DBS concept by AI Foundations, with attribution. Feedback welcome.
More from coding & agent
- Give Claude real designer assets: 40 vetted resources to fix vibe-coded UI, says founder — PrajwalTomar_ · 2026-10-05
- Claude Opus 5.5 builds an interactive lens lab in one 1h26m shot for $25.66 — nikola_mr64990 · 2026-10-05
- Andrew Ng Releases Free 1-Hour AI Agent Course: From Fundamentals to Multi-Agent Systems — nikola_mr64990 · 2026-10-05
- 15 Free Claude Courses Curated: Claude Code, RAG, MCP and Anthropic Official Training — nikola_mr64990 · 2026-10-05
- Same prompt, three tools: ChatGPT vs Claude vs Figma Make compared — Tegadesigns · 2026-10-05
- OntoPrune cuts 83% of context tokens, speeding local CPU TTFT by 6.7x — vigmarcarlo · 2026-10-05