メニュー
Keep Specrew's Copilot launch flags aligned with platform-specific REPL behavior
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Run live continuous co-review or replay persisted reviewer evidence for an iteration.
Run live continuous co-review or replay persisted reviewer evidence for an iteration.
Run live continuous co-review or replay persisted reviewer evidence for an iteration.
Perform a Specrew human-verdict boundary stop on the Claude host. Renders the FULL Rule 46 six-section human re-entry packet AND the verdict options as one Markdown message, with the AskUserQuestion picker disabled so the packet cannot collapse into the picker's short header/option fields. Invoke at EVERY human-judgment boundary stop (specify, clarify, plan, tasks, before-implement, implement, review, retro, feature-closeout, lifecycle-end). Triggers: boundary stop, verdict, approve / redirect / send back, why I stopped, human re-entry packet, gate stop.
{what this skill teaches agents}
Run Specrew's per-lens design workshop and collaborative design-analysis. Use whenever you work the design lenses for a feature: at specify/intake (the lens workshop) and at the design-analysis stop (co-design the architecture), and RE-INVOKE at the start of EACH new lens (architecture, data, ui-ux, security, integration, devops, requirements/NFR, observability, component). Triggers: design, design lens, lens workshop, design-analysis, architecture, trade-offs, co-design, explore options, decompose, or moving from one lens to the next. Tells you to facilitate each lens as a discussion, surface diagrams the human can actually SEE (console ASCII inline; mermaid/html to a file with a clickable file:/// link), co-design components/responsibilities/flows WITH the human instead of handing over finished options, capture the agreements, and which per-lens md to load.
| name | copilot-launch-contract-divergence |
| description | Keep Specrew's Copilot launch flags aligned with platform-specific REPL behavior |
| domain | runtime |
| confidence | high |
| source | earned |
| tools | [{"name":"powershell","description":"Run the existing start-command integration coverage after changing launch flags or messaging","when":"When `scripts\\specrew-start.ps1` changes Copilot CLI argument composition"},{"name":"view","description":"Compare launch code, integration assertions, and test evidence","when":"When platform-specific runtime behavior must stay documented and testable"}] |
Use this when specrew start changes how it launches Copilot across Windows and Linux/macOS. The runtime contract is not symmetric: the same Copilot CLI flags can keep Windows interactive while breaking Linux/macOS, so launch code, user messaging, tests, and evidence must move together.
-i bootstrap auto-load on every platform.--autopilot behavior unchanged and mutually exclusive with --mode interactive.--mode interactive; add --allow-all only when explicitly allowed.--agent, --add-dir, and -i without --mode interactive or --allow-all, and warn users to expect approval prompts for bootstrap file reads.scripts\\specrew-start.ps1, tests\\integration\\start-command.ps1, and specs\\019-specrew-distribution-module\\test-evidence\\us5-cross-platform.md together.--mode interactive to hold the REPL open, while Linux/macOS dropped it because Copilot CLI v1.0.48 already defaults the -i flow into the REPL there and exits one-shot when --mode interactive is added.pwsh -NoProfile -File tests\\integration\\start-command.ps1 is the regression proof for Windows-side launch composition after these changes.copilot --agent ... -i flags.--allow-all or --mode interactive on Linux/macOS without fresh live evidence that the CLI behavior changed.