Beads/Todo/Rulesync Beads Skill Abstraction
Rulesync Beads skill abstraction
- In Beads/Todo: this page is filed here as a reminder to capture one portable story for Beads /
[[bd]] across Agentic Engineering Tools, instead of re-deriving setup from each vendor doc every time. Problem
- Official docs describe per-IDE / per-agent wiring so agents know how to use
bd (hooks, plugins, CLI setup, etc.). Beads — IDE setup is the hub for that. - This garden already uses
[[rulesync]] as the single source for rules, commands, skills, and related AI-tool config that fans out to Claude/Code, CursorAI/CLI, Codex/CLI, Copilot-style targets, and the rest. - Gap: there is no rulesync-authored rule or skill yet that tells every tool the same baseline: this repo uses Beads; run these setup or registration steps for your runtime; keep issue state in sync the same way.
Goal
- Add something under
.rulesync/rules/ and/or .rulesync/skills/ that:- Points agents at
bd when the repo has .beads/ (or documents how to detect “Beads project”). - Summarizes or links the IDE-setup matrix (what
bd setup … exists, what needs a plugin, what is hook-only) without duplicating upstream forever. - Aligns with existing workflow rules—e.g.
[[AI/Coding/Concept/Land the Plane]] / .rulesync landing-the-plane material that already expects agents to run bd sync before push where Beads is in use.
- Optional stretch: a progressive-disclosure skill (see
[[rulesync/How To/Write Skills With Progressive Disclosure]]) so lightweight installs only see a checklist, and deep references stay in references/.
Prior art in this graph
Done when
- A generated rule/skill lands in the targets you care about (Cursor, Claude Code, Codex CLI, …) from rulesync, and agents reliably register or reuse Beads per IDE setup without one-off instructions in each session.
Related