Odin Run

Use when the user invokes /duet, says pair on this, or faces aesthetic, architectural, or irreversible decisions. Not for one-shot fork resolution: use decide.

Use when implementing from a plan, spec, clear build request, or settled ticket, orchestrated or standalone. Don't use for open-ended debugging (figure-it-out), exploration, or read-only research.

Use when asked to generate a bash wizard for provisioning, credentials, dashboards, migrations, or cutovers a human performs. Not for remote, credential, publish, deploy, or irreversible changes.

Use when a user starts a new work session and asks to split human decisions from agent execution. Not for remote, credential, publish, deploy, or irreversible changes.

Use when asked to set, list, pause, update, or delete a reminder or local task that fires at a confirmed time or interval. Not for destructive, remote, credential, publish, or deploy operations.

Use when monitoring a file, log, endpoint, or artifact for drift or errors, or polling a target until a predicate holds. Modes: watch (default), until. Not for source or remote-system changes.

Use when a build, QA pass, demo, user complaint, or abandoned attempt leaves the next pass needing lessons rather than code. Not for changelog extraction. Not for session handoff: use handoff.

Use when one issue or PR is too large for a worker and partitions into disjoint write sets. Not for independent parallel tasks: use parallel-launch. No remote or irreversible changes.

Use when a run could be over- or under-powered, before dispatching a subagent, or the user asks how hard to think about a task. Recommends one tier and effort level. Not for executing the sized task.

Use when non-trivial work should run a matched playbook to verified real-surface completion, or a bespoke workflow when none fits. Not for remote, credential, publish, deploy, or irreversible changes.

Use when the user says ship it now, no phases, or finish it now, or would defer work to phases, rollouts, follow-up PRs, or TODOs. Not for remote, credential, publish, deploy, or irreversible changes.

Use when a button click, dashboard check, or human verdict sits inside an iteration loop and needs an autonomous gate. Not for remote, credential, publish, deploy, or irreversible changes.

Use when the user explicitly asks to create, update, query, or archive Notion pages, databases, or blocks. Not for bulk operations, recursive edits, or widening the workspace scope.

Use when the user explicitly asks to share a message on Slack through an incoming webhook. Not for posting without an explicit request or for retrying an ambiguous delivery.

Use when a user opens a build cycle in a repo with an iteration convention and wants the casebook seeded with real content. Not for ongoing triage or multi-repo setup.

Use when a completed agent run must be recorded as durable, queryable evidence. Not for remote, credential, publish, deploy, or irreversible changes.

Use when asked to run /artifact-arena to generate and judge competing artifact implementations. Not for remote, credential, publish, deploy, or irreversible changes.

Use when asked to /feedback-sweep [setup|reconfigure] [mode:non-interactive]. Not for pattern discovery: use feedback-pattern-sweep.

Use when a ticket DAG from a complete specification needs parallel execution into a green draft PR. Not for a settled ticket with no spec: use work. Not for ticket decomposition: use to-tickets.

Use when the caller supplies a falsifiable out-of-happy-path invariant and a finite budget. Restores via bounded patches or reverts. Not for normal feature delivery or universal retries.

Use when work splits into independent sub-tasks or cross-domain research. Not for durable scope ownership with per-scope review: use partition-scopes-to-subagents.

Use when a user says execute with subagents or hands over an ordered multi-task plan. Not for credential, paid, or deployment work, or any mutation outside the skill's designated workspace.

Use when a scheduled or watcher tick fires and a lightweight pulse must capture current external state. Not for remote, credential, publish, deploy, or irreversible changes.

Use when work decomposes across subagents or phases: audits, migrations, research sweeps, or scale one context cannot hold. Not for remote, credential, publish, deploy, or irreversible changes.

Use when the user invokes it to append a structured decision record to an append-only TSV log, ending with an Attention section for reviewers. Not for an ephemeral visual: use show-me.

Use when a long agentic project needs each cycle to end in a learning memo and a keep/iterate/restart decision. Not for single-pass builds.

Use when loop scaffold files have drifted from their provenance-pinned templates. Not for remote, credential, publish, deploy, or irreversible changes.

Use when guardrail-adjacent material (stealth, scraping, privacy, IP, policy, security) or mixed reversible/irreversible, sensitive work, or the user says "autobahn this". Not for carved risky work.

Use when a configured loop misbehaves, produces unexpected results, or its setup soundness is questioned. Not for tasks that require source or remote-system changes, not for loop design.

Use when a task needs gate-file proof before a done claim. Also fires on "gate-file-completion", "depth tree", "gates", or "do not stop until it is done".

Use when asked to evaluate a proposed commit, merge, or auto-merge against a gate config. Not for landing PRs: use gate-and-merge. Read-only.

Use when asked to draft copy-ready /goal objectives for long-running agents. Not for source or remote-system changes.

Use when a high-priority run reaches at least 90% of its budget and requests an extension. Never self-grants. Not for low-priority runs: let those exhaust silently.

Use when the user wants to classify a surface state and page an on-call API when triggered. Not for read-only anomaly watching without paging: use watch-for.

Use when work decomposes across subagents or role panels and needs coupling-based orchestration. Not for cloud-agent task-graph orchestration: use cloud-task-orchestrator.

Use when a proven watch pattern should become a reusable harness artifact with configurable inputs. Not for remote, credential, publish, deploy, or irreversible changes.

Use when a human explicitly runs /cloud-task-orchestrator for a large task across cloud agents to drain a verified task graph. Not for local subagent coordination: use orchestration-patterns.

Use when asked to change state in an external system: propose the write locally and halt at the human gate without executing. Not for remote, credential, publish, deploy, or irreversible changes.

Use when the user asks to reply to or follow up on a specific Slack thread. Not for top-level messages, deleting messages, or composing the reply text.

Use when asked to partition non-overlapping scopes across subagents and drive each to completion. Not for concurrent independent sub-tasks: use parallel-launch.

Use when a candidate patch or answer needs independent corroboration before it is trusted. Not for multi-stance investigation: use council.