Lucas WintherandClaude Opus 5.5
15740a054f
Adds AGENTS.md (clean-room and Bun rules), pointer files, agent permission configs, docs tree, the STaurtomation design spec (draft) and sanitized licensing/competitor research. Co-Authored-By: Claude Opus 5.5 <[email protected]>
3.3 KiB
Decisions
Append-only. Never edit or delete an entry — supersede it with a new one that links back.
One entry per decision a future maintainer would otherwise have to reverse-engineer.
No index: grep -n '^## ' docs/DECISIONS.md lists every entry (date, title, line number).
2026-10-30 — Release 1 is homelab-ready, without the community-node compatibility layer
Status: accepted Context: Reimplementing n8n's node-facing interface is the project's highest legal-risk area. Decision: Release 1 targets single-box homelab installs (spec §10). The n8n importer ships as best effort with clear warnings; the community-node layer ships only after legal review. Alternatives: Waiting for legal review before any release (delays everything). Consequences: Imported workflows that use community nodes get placeholders until Release 3.
2026-10-30 — AI agents may build every area, behind mandatory guards
Status: accepted Context: The project is built mainly by AI agents, whose training data likely includes the public source of the products being superseded. Decision: Allowed, with CI similarity scanning, per-PR provenance notes, a dependency denylist, agent tooling restrictions and human sign-off on importer/compat work (spec §3.2 rule 5a). Alternatives: Human-only implementation of high-risk areas (not staffable). Consequences: CI must carry the similarity and denylist guards before high-risk work starts.
2026-10-29 — License AGPL-3.0-or-later with a CLA
Status: accepted Context: Owner wants the project to stay open while retaining copyright and relicensing freedom. Decision: AGPL-3.0-or-later for everything; outside contributions require a CLA. Alternatives: MIT (allows closed forks); GPL-3.0 (hosted-modification gap); DCO only (no relicensing rights). Consequences: Only AGPL-compatible dependencies; CLA tooling needed before accepting outside PRs.
2026-10-29 — Bun-native stack, React editor, Better Auth
Status: accepted
Context: n8n community nodes are TypeScript/npm packages; owner wants Bun + TypeScript.
Decision: Bun with Bun.serve() and Bun built-ins (per bun init conventions); React + React
Flow + shadcn/ui + Tailwind via Bun HTML imports; Better Auth for authentication, with Passport as
fallback if its SAML/SCIM support proves immature.
Alternatives: Next.js (redundant server, public-site oriented); Vite and Hono (superseded by Bun
built-ins); Passport first (Express-centric).
Consequences: Verify Better Auth on Bun before building auth (spec §12).
2026-10-29 — Adopt AGENTS.md as single source of truth
Status: accepted
Context: Agents had no durable project context and no consistent workflow between sessions.
Decision: AGENTS.md is canonical; CLAUDE.md/GEMINI.md are pointers. Specs are permanent in
docs/specs/. Plans live in docs/plans/ only while work is in progress and are removed before the
merge request. Both override the superpowers default paths.
Alternatives: Per-tool rule files (drift, synchronisation debt); no rules file (agents re-explore
the repo every session — measurably slower and more tokens).
Consequences: Rule changes happen in one file; pointer files must stay thin. Plans are working
artefacts: kept after merge they go stale and contradict the code, so the spec is the record.