mkXultra

Orchestrate a multi-model build-and-review workflow. Use when Codex receives YAML input with a builder model, optional builder session_id, fixed reviewer sessions, a shared review prompt, a reviewer verification prompt, done criteria, and a max loop count, and must run a builder, collect review findings via files, merge and deduplicate findings, resolve contradictions by resuming reviewer sessions, loop the builder until done, and return each agent's final model, session_id, and status.

Orchestrate a multi-model review-and-fix workflow using only ai-cli run and ai-cli wait. Use when Codex receives YAML input with fixed reviewer sessions, a shared review prompt, and a required fixer model/session, and must run codex-ultra, gemini-ultra, and claude-ultra reviews, enforce structured reviewer output, merge and deduplicate findings, resolve contradictions by resuming reviewer sessions, optionally run a fixer, and return each agent's final model, session, and status.

Orchestrate a multi-model review-and-fix workflow using only mcp__acm__run and mcp__acm__wait. Use when Codex receives YAML input with fixed reviewer sessions, a shared review prompt, and a required fixer model/session, and must run codex-ultra, gemini-ultra, and claude-ultra reviews, enforce structured reviewer output, merge and deduplicate findings, resolve contradictions by resuming reviewer sessions, optionally run a fixer, and return each agent's final model, session, and status.

Orchestrate a 3-model parallel review workflow using mcp__acm__run and mcp__acm__wait. Runs codex-ultra, gemini-ultra, and claude-ultra reviews in parallel, enforces structured output, merges and deduplicates findings, resolves contradictions, and returns a final merged review report. No fixer agent is involved.

Run the coding taskforce buddy workflow: implement with codex-ultra, run unit tests with gpt-5.1-codex-mini, triage results, and iterate fixes using the same session_id. Use when asked to follow taskforce_coding_buddy.md or to orchestrate this multi-step coding workflow.

Run the taskforce coding team workflow: execute the taskforce-coding-buddy skill with codex-ultra, execute the taskforce-review-report skill with gemini-ultra, and loop based on the report STATUS. Use when asked to follow taskforce_coding_team.md or to orchestrate this coding+review loop.

Run the taskforce review workflow: perform AI code reviews with mcp__acm__run (codex-ultra, add gemini-ultra if multiple reviews are needed) and then produce a review report with STATUS and DESCRIPTION. Use when asked to follow taskforce_review_report.md or to orchestrate review + report.

PRのレビューコメントを取得してタスク化する。レビュー対応やPRのコメント確認が必要な時に使用

Re-enter mew planning or long-session work after context compression, interruption, or a user asks what to do next. Use to recover current work, durable product decisions, deferred decisions, and the next safe action from mew's own state and memory.

Run isolated mew side-project implementation as mew-first dogfood, update SIDE_PROJECT_ROADMAP_STATUS.md, and record structured M6.13.2 telemetry for M6.16 implementation-lane polish.

Evaluate mew as a passive AI product and as a shell/body that an AI model might want to inhabit. Use when asked whether mew is good enough, whether it feels usable as the next execution form, or what should improve next.

Update mew roadmap milestone status. Use when asked to assess progress against ROADMAP.md, refresh milestone statuses, decide the next roadmap target, or preserve roadmap progress across context compression.

Run bounded mew roadmap/coding tasks with mew as the first implementer and Codex as reviewer. Use when implementing mew tasks after M6.7, assessing mew-first success or failure, deciding whether to repair loop substrate immediately, or preserving autonomy-credit accounting across context compression.

Decide what to do immediately after a git commit in the mew repo. Use when a commit just landed and you must choose whether to report to the user or continue working under a long-session output gate.

Save durable mew context before a long session, context compression risk, interruption, or major handoff. Use to write current intent, next action, validation state, blockers, active work-session notes, and durable product decisions into mew's own memory/state before continuing.

Adversarially review proposed mew self-improvement loops for product-goal drift, M5 safety boundaries, weak evidence, missing verification, hidden rescue edits, and whether the loop should be approved, rejected, or revised.