What problem does it solve? After an epic is planned, deciding which stories to execute together — and in what order — is judgment work that usually lives in someone's head. This Skill reads an epic's not-done stories and paired testplans, then produces a reviewed-before-acting Implementation Strategy: stories grouped into coherent chats with lanes, sub-agents, paste-ready triggers, and verify commands, plus an unattended autopilot track. ## Core Features & Use Cases - Judgment-led chat grouping: Groups stories into chats of 2–5 based on shared domain, dependencies, and file overlap, with a written rationale per chat and soft-bound deviations recorded. - Dual output artefacts: Writes a human-readable markdown report and a structured JSON sidecar under 41-Reports/execution-strategy/ that the dashboard's Implementation Strategy view renders. - Autopilot track: Always regroups the same stories into unattended runs of at most 10 chats, each with risk-ordered stories, checkpoints, stop conditions, halt blast radius, and a self-contained paste-ready prompt. - Dry-run safety: Never modifies story status; it only reads frontmatter and writes the two report artefacts, with structural shape tests validating each emission. - Use Case: Run /tandem:execution-strategist EPIC-100 after planning an epic to get a phased execution plan you review, then hand off chats to /tandem:execute-batch. ## Quick Start Ask the assistant to plan execution for an epic, for example: run the execution-strategist on EPIC-100 to group its stories into chats and produce the Implementation Strategy report.
