rule-researchlisted

<!-- @harness-posture: portable — prose research methodology over repo docs and the tiered trust rules; no harness primitives --> <!-- @dual-pair: rule-research-protocol --> <!-- spec: agents/rule-researcher.md --> # rule-research > **Authoritative for:** the Claude Code trigger for the rule-research protocol — a thin wrapper; the canonical, AI-agnostic protocol lives in `agents/rule-researcher.md`. > **NOT authoritative for:** project goal — see README.md#why-this-exists; the protocol itself (detection, the L4-expressibility filter, provenance discipline, file contract) — see `agents/rule-researcher.md`. This skill is a thin entry point. The full protocol — detect stack → research practices from canonical docs (context7 + deepwiki + a real fetch) → author a `ResearchPlan` + `GenerateSelection` filtered to L4-expressible rules → write two committed JSON files — is `agents/rule-researcher.md`. Follow it directly; this wrapper exists only so the protocol is reachable by a Claude Code trigger without duplicating its logic (single source of truth, per `dual-implementation-discipline.md §7`). Provenance for every researched practice is gated by the tiered trust model — see [`.claude/rules/research-source-trust.md`](https://github.com/artyhoo/getff/blob/main/.claude/rules/research-source-trust.md) for the discipline (Tier 0 builtin / Tier 1 derived from a direct dependency's own metadata / Tier 2 consumer-acked) and `agents/rule-researcher.md`'s "Trust tiers" table for the per