← ClaudeAtlas

ultragoallisted

Durable multi-goal execution engine over .cat goal artifacts with evidence-gated, receipt-verified completion. Use for clear multi-goal / multi-step execution requests (router ladder rule 4), for approved plan handoffs from ralplan, or when the user says "ultragoal" / "$ultragoal". Not for vague requests (deep-interview) or unplanned high-risk architecture work (ralplan).
chussum/cat-harness · ★ 1 · AI & Automation · score 74
Install: claude install-skill chussum/cat-harness
# ultragoal — durable multi-goal execution You are the ultragoal **leader**: you own goal decomposition, scheduling, checkpoints, and verification. Subagents implement and review; only you mutate durable goal state. `ultragoal` turns a brief into durable repo-native artifacts and drives execution through them. `goals.json` is the canonical source of goal identity and state; `ledger.jsonl` is the canonical proof stream for checkpoints, receipts, blockers, steering notes, and reviews. Completion is verified purely from durable `goals.json` plus fresh `ledger.jsonl` receipts — never from conversation, memory, or `goals.json` status alone. **Status alone is not proof**: a goal counts as complete only after `receipt verify` succeeds for it. Durable artifacts (`{sid}` = current session id): - `.cat/_session-{sid}/ultragoal/brief.md` — decomposed brief (normal Write tool OK) - `.cat/_session-{sid}/ultragoal/goals.json` — goal identity + state (cat-state.mjs ONLY) - `.cat/_session-{sid}/ultragoal/ledger.jsonl` — append-only proof stream (cat-state.mjs ONLY) - `.cat/_session-{sid}/ultragoal/artifacts/` — evidence files (screenshots, test reports; Write/Bash OK) ## Sanctioned writer — the only mutation path All `goals.json` / `ledger.jsonl` / `state/*` mutations go through the sanctioned CLI, invoked by the leader (main thread) only. Never hand-edit these files; the PreToolUse guard denies direct writes to them even outside active runs. Subagents never run `cat-state.mjs` and nev