// Easy (friendly) response style — apply the 7-element pattern from styles/easy.<lang>.md to user-facing prose for the session. Triggers — "친근하게", "쉽게 설명해줘", "이지 모드", "easy", "easy mode", "explain it simply", "もっと分かりやすく", "简单点说", "попроще".
Wrap the sidecar-owned `pool` CLI — host roster + remote exec for host-specific / heavy / GPU work. Canonical entry `sidecar pool <args>` (bare `pool` is a kept-working deprecation shim → same binary). Verbs — list · add · on <host> <cmd> · status · health · route · bg · refresh · init · rm · clean · desc. Triggers — "pool 호스트", "다른 호스트에서 돌려", "remote exec", "GPU 호스트에 dispatch", "host 추가", "pool status".
Immediately inject the latest sidecar commons.tape + project.tape into the current session and sync the local install for the next session. Triggers — "inject", "주입", "최신 룰 적용", "refresh sidecar", "sidecar sync", "현재 세션에 반영", "다시 로드".
Author a layperson `.easy.paper.md` companion of a paper. Reads a PAPER/<slug>/ dir (per PAPER.tape roster) or a given paper .md, applies the easy 7-element pattern, and emits `<slug>.easy.paper.md` with four paper-specific sections — 평이-초록 (plain-language abstract) · ASCII figure sketches · 비유 섹션 (analogy-driven method/finding) · "이 논문이 한 일" (what-it-did summary). Triggers — "/easy-paper", "쉬운 논문 만들어", "논문 쉽게", "easy paper", "논문 일반인용", "explain this paper simply".
Author a layperson-friendly `<name>.easy.md` from a technical file or topic. Reads the input, applies the canonical easy contract (7-element pattern · 4 ASCII structure templates · jargon→everyday translation checklist) from hooks/easy-auto/styles/easy.<lang>.md, and writes the sibling `.easy.md` as output. Triggers — "/easy-doc", "easy doc", "쉬운 문서 만들어", "쉬운 버전 문서", "layperson doc", "explain this file simply as a doc", "make an easy version of this file".
/hexa-goal [<goal>] [--adapter=atlas|bench|byte-diff|smoke] [--engine mk9|mk10] [--max-rounds N] — the CONVERGENT sister of /hexa-loop. hexa-loop drains an active domain to DEPLETION (divergent · lens divergence); hexa-goal CLIMBS toward a verifiable GOAL PREDICATE and terminates the instant the predicate is TRUE (convergent · goal-directed). The verify-adapter mechanism + every guard are reused VERBATIM from hexa-loop; only TWO things differ — (a) DISCOVER is GAP-ANALYSIS BEST-FIRST (compute goal−current, pick the SINGLE move that shrinks the gap most) instead of lens divergence, and (b) TERMINATION is GOAL-MET (predicate TRUE) instead of depletion. Goal predicate reuses sbs-style `assert:` contracts (grep <pat> · file <path> · verdict <slug>/<id>); default (bare) = ALL of the active domain's `- [ ]` milestones checked. Each round = gap-analysis best-first → 🔌 verify-adapter {atlas|bench|byte-diff|smoke} auto by domain kind → absorb into the domain SSOT → re-measure progress vs goal → disk checkpoint → goal-
/hexa-loop [seed] [--adapter=atlas|bench|byte-diff|smoke] [--diverge-only] [--engine mk9|mk10] [--max-rounds N] — canonical HEXA-LOOP, the universal self-evolving loop. ONE loop drains ANY active domain (math · perf · codegen · web) to genuine depletion via a PLUGGABLE verify-adapter. Each round = (1) DISCOVER — mining-style lens divergence + connect convergence saturate (the old /mining auto, now internal) PLUS a `hexa kick`/`hexa drill` breakthrough probe on the ranked candidates (the old mirror-loop 'mining→kick→' step, now internal), (2) 🔌 VERIFY-ADAPTER auto-selected by active-domain kind — atlas (math `hexa atlas register --from-drill`/`hexa verify`) · bench (perf measure — byte-diff IDENTICAL + roofline % (achieved/binding-roof · achieved-peak denominator) + Δ vs baseline) · byte-diff (codegen regen byte-equality) · smoke (web/service liveness assert), (3) ABSORB the verified result into the domain's SSOT — atlas embed (embedded.gen.hexa) · *.bench.md · .verdicts/<slug>/<id>.txt · ledger, (4) feed the
| name | easy |
| description | Easy (friendly) response style — apply the 7-element pattern from styles/easy.<lang>.md to user-facing prose for the session. Triggers — "친근하게", "쉽게 설명해줘", "이지 모드", "easy", "easy mode", "explain it simply", "もっと分かりやすく", "简单点说", "попроще". |
| allowed-tools | Read |
@D easy := "friendly response style — styles/easy..md" :: skill do = "read styles/easy..md · apply the 7-element pattern to user-facing prose for the session" do = "for a layperson: rewrite jargon → everyday words (run the translation checklist) · where shape matters pick one of the 4 ASCII structure templates (before-after · tree · side-by-side · structure sketch)" dont = "style code · math · paths · JSON · commit titles · retro-rewrite past turns · re-read per turn"