菜单
Create a host-native `/goal` prompt for bounded ORCA plan work, especially when the user wants an agent to keep working through ready ORCA plans, delegate blocker fixes to subagents, record true blockers, and move on without stalling.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Implement planned changes with scoped edits, verification, and respect for existing project patterns.
Create bounded delegation briefs and structured executor returns for ORCA Framework workflows.
Prepare release readiness checks, release notes, rollback guidance, and handoff.
Record time, retries, and optional token, cache, and cost metrics for ORCA Framework workflows without fabricating precision.
Provide lightweight, skill-sensitive coaching without slowing down normal ORCA Framework execution.
Decide when risky work must pause for approval and record the approval request clearly.
| name | orca-goal-prompt |
| description | Create a host-native `/goal` prompt for bounded ORCA plan work, especially when the user wants an agent to keep working through ready ORCA plans, delegate blocker fixes to subagents, record true blockers, and move on without stalling. |
Generate a copy-ready native /goal prompt that turns a predefined chunk of ORCA work into a bounded execution loop.
Use this after ORCA has ready plans, issues, or milestones. If the chunk is not explicit, ask the user whether to use the default chunk: all ready ORCA plans.
/goal prompt focused only on that blocker.Return one copy-ready prompt, not an essay:
/goal Work through [CHUNK] for [PROJECT] until every item is completed, recorded as truly blocked, or intentionally skipped with a reason.
Start by inspecting [PLAN SOURCES], current repo state, existing worktree/branch policy, and [TRACKER]. Preserve unrelated dirty work.
Loop:
1. Pick the next ready unblocked item from [CHUNK].
2. Implement the smallest production-ready slice that advances it.
3. Run the cheapest meaningful verification for that slice.
4. Update [TRACKER] with status, files changed, verification, blockers, and next step.
5. If blocked, collect exact blocker proof. If a subagent could reasonably unblock it, launch this subagent goal and keep working on another independent item:
/goal Fix blocker: [BLOCKER]. Work only on [OWNED SCOPE]. Use [ARTIFACTS]. Return status, files changed, verification, exact blocker proof if still blocked, and recommended next step. Do not touch unrelated work.
6. When a subagent returns, ingest only verified results, reconcile conflicts, update [TRACKER], and continue.
7. If blocker is external, repeated, or needs user credentials/approval, record it in [BLOCKER REPORT TARGET] with exact command/output/artifact path and move on.
Stop when [CHUNK] has no remaining actionable work, verification proves done, user approval is required, or all remaining work is truly blocked with fresh evidence.
Final answer: completed items, blocked items with proof, skipped items with reason, verification run, branch/PR state, and next human action.
.orca/work-items/**, docs/** plan/status files, checked-in ORCA plans, and current issue board.templates/blocker-report.md shape.The generated prompt must make continuation, delegation, blocker recording, verification, and stop conditions explicit. It must not imply infinite work.
Use with orca-goal-mode, orca-background-mode, orca-agent-orchestration, orca-delegation, and orca-receipts.