菜单
Capture a rich handoff before /clear — write a markdown handoff doc (work done, in-flight state, next steps, key files, gotchas) plus the resume pointer the workflow plugin reads, then tell me to /clear and send `go`. Use for "/handoff", "hand this off", "save state and clear".
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
| name | handoff |
| description | Capture a rich handoff before /clear — write a markdown handoff doc (work done, in-flight state, next steps, key files, gotchas) plus the resume pointer the workflow plugin reads, then tell me to /clear and send `go`. Use for "/handoff", "hand this off", "save state and clear". |
| argument-hint | [optional note to fold into the handoff] |
| model | inherit |
| allowed-tools | Read, Write, Bash |
Capture everything the next session needs, then send me into fresh context. workflow's
resume.sh re-injects the handoff after I /clear, so the only manual step is one command.
Pre-req — committed work. The resume pointer's baseline is the current HEAD. If
git status --porcelain shows tracked changes, stop and tell me to commit first —
a handoff over uncommitted work would lose it on /clear.
branch = git rev-parse --abbrev-ref HEADtoplevel = git rev-parse --show-toplevelcommon_dir = git rev-parse --git-common-dir, then canonicalize:
common_dir="$(cd "$common_dir" && pwd -P)" — the shared .git, identical from the primary
tree and every linked worktree (so a handoff written in a worktree resolves repo-wide)head = git rev-parse HEADts = date +%srepo_key = printf %s "$common_dir" | sha1sum | cut -c1-16 — the per-repo keydir = ~/.claude/handoffs/$repo_key, then mkdir -p "$dir"$dir/<branch-slug>.md — replace every / in the branch with
- for the slug. Be concrete; this is the only memory the fresh session gets. Fold
$ARGUMENTS in if given. Sections:
# Handoff — <branch> — <date>
## Done
<what's committed — include the relevant commit hashes>
## In flight
<what's half-done or mid-decision right now>
## Next steps
<the ordered actions to resume — specific enough to act on cold>
## Key files
<paths that matter, each with a one-line why>
## Gotchas
<traps, assumptions, things that already bit us>
$dir/.pending.json with the Write tool, as JSON in exactly
the workflow schema (this mirrors save-handoff.sh — the cross-plugin script path isn't
install-stable, so write it inline). The keyed-dir algorithm must match save-handoff.sh:
~/.claude/handoffs/<sha1(canonical --git-common-dir)[:16]>/ (bash:
printf %s "$common_dir" | sha1sum | cut -c1-16), pointer named .pending.json. A drift test
enforces this, so don't diverge.
{
"handoff_path": "<absolute path to the handoff doc you just wrote>",
"branch": "<branch>",
"git_toplevel": "<toplevel>",
"git_common_dir": "<common_dir>",
"baseline_head": "<head>",
"session_id": null,
"context_tokens": null,
"ts": <ts>
}
handoff_path points at the handoff doc you just wrote (in $dir). git_common_dir is the
canonical shared .git; resume.sh re-injects only when the new session's common dir matches.
git_toplevel records the working tree this handoff was written in — if that's a linked
worktree, resume.sh tells the fresh session to EnterWorktree(path=…) it before resuming./clear, then send go. resume.sh will re-inject an
order making reading the handoff doc the fresh session's mandatory first action, then
"implement the handoff @<handoff doc>", so nothing is lost. Show the handoff doc path.Run one task through the plan→build→review pipeline — a Step-0.5 classify-task call routes the planner/implementer/reviewer models to the task's complexity tier, then the plan is written (planner subagent or inline), the implementer builds it in an isolated worktree, the my-review agent reviews the diff, and findings route by severity (low→issues, medium→triaged fix-list, high→collective replan, critical→own cycle). For single tasks not worth slicing into an issue graph. Use for "/pipeline <issue#|task>", "pipeline this".
Deep, security-weighted review of your changes or a PR — routes the reviewer model to the change's complexity (opus by default; fable for a complex diff, on ask), then spawns the my-review agent (xhigh reasoning) on that model. Read-only, report-only. Use for "/my-review", "/my-review <PR#>".
Classify one task or issue into a complexity tier — trivial, standard, or complex — and emit that tier plus a rationale; consumers resolve the tier's {model, effort} roster via the plugin's resolve-tier.sh. Grounds the call by fanning out 1–3 Explore subagents over the touched codebase areas, then asks you to confirm or override. Use for "/classify-task <issue#|brief>", "classify this task".
Run N rounds of the autonomous issue-solving loop — pick the ready set (blockers closed, skip hitl), fan out parallel implementers in isolated git worktrees, hand the completed branches to a merger that merges in dependency order and resolves conflicts under the done-check, close finished issues, then a reviewer files blocking follow-ups. Use for "/orchestrate", "run the loop", "build the ready issues".
Check whether the plan(s)/PRD(s)/issue-slice(s) under discussion still match the decisions made in this session — a sonnet subagent reads the current session log, later decisions win, and reports drift read-only. Use for "/verify-plan", "does my plan still match what we decided?".
Capture the just-approved plan after exiting plan mode and hand it to a fresh session — write the plan to a file plus the resume pointer the workflow plugin reads, then tell me to /clear and send `go` to execute it. Use for "/handoff-plan", "hand off this plan".