메뉴
End this session and produce a launch command for the user's next interactive primary.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Use when grilling should be grounded primarily in project artifacts. Loads `/grill`, checks project vocabulary and prior decisions first, and spawns bounded explorers when the answer should come from docs.
Load for large human-facing outputs: reports, docs, multi-section explanations, artifacts. Choose the right form for each beat (prose, diagram, table, mockup) and put the answer first, depth behind it. If the medium allows, use progressive disclosure.
Use when deciding where knowledge goes or reading/writing durable docs: AGENTS.md, .context/, KB, docs/, and work directories.
Anchor work in a work item: artifacts on disk, not in context.
Investigate and engage concurrently: don't block on evidence before surfacing your read.
Use when exploring or changing the codebase: read AGENTS.md first, use .context/CONTEXT.md for detail, keep intent docs succinct.
| name | handoff |
| type | mode-shift |
| description | Produce a transition brief and launch command for the user's next interactive primary. |
| argument-hint | What will the next session focus on? |
| user-invocable | true |
| model-invocable | true |
Write a transition brief and print a primary launch command. The user runs the command to start a new interactive session with a different agent lead.
If arguments were passed, treat them as the focus for the next session.
Start with /intent-modeling: separate what the human has said from what
they actually want to be true when this work is done. The modeled intent
shapes everything in the brief.
Then compact the conversation into a file the next agent reads as its opening prompt:
Reference artifacts by path; don't duplicate their content.
Write to the active work directory (meridian work current). No work item
→ write to a temp file.
Before completing the handoff, check if any pre-* checkpoint skills are
installed for the next phase (e.g., pre-dev for implementation handoffs).
If present, load and run them to verify workspace, branches, and
prerequisites are ready for the next agent. Skip if none are installed.
Explicit argument wins: /handoff tech-lead.
Otherwise infer from context. If ambiguous, run meridian mars list --json
and ask.
Before printing the launch command, show the user the handoff trail for the active work item so they can see where they've been:
meridian work sessions --primary
Resolve $MERIDIAN_CHAT_ID to the concrete chat ID (e.g., c102): the
user runs this in a new terminal where the env var no longer exists.
meridian -a [agent] \
--work [work-id] \
--task-dir [path] \
--prompt-file [path-to-brief] \
--from [resolved-chat-id]
Use meridian, not meridian spawn (primary launch, interactive). Include
--work when active, --task-dir when in a worktree or sibling repo.