菜单
Fast path to apply a targeted fix without running the full flywheel.
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
Relentless goal/design interview that sharpens framing and writes durable docs (brainstorm artifact, ADRs, glossary) as decisions crystallize. Use when refining a flywheel goal, pressure-testing scope, or "grill with docs".
Start or resume the full agentic coding flywheel. Drives the complete workflow: scan → discover → plan → implement → review.
Set up flywheel prerequisites for this project.
One-shot diagnostic of every flywheel dependency — MCP connectivity, Agent Mail liveness, br/bv/ntm/cm binaries, node version, git status, dist-drift, orphaned worktrees, and checkpoint validity. Use when debugging toolchain issues, before starting a new session, after /flywheel-cleanup, or as a CI gate.
Strategic gap analysis between vision (AGENTS.md / README.md / plan docs) and what's actually implemented. Converts gaps into beads, optionally launches a swarm. Use when "reality check", "where are we really", "gap analysis", "did we drift", or before declaring a long-running project done.
Launch a parallel swarm of agents to implement multiple beads simultaneously.
| name | flywheel-fix |
| description | Fast path to apply a targeted fix without running the full flywheel. |
Apply a targeted fix: $ARGUMENTS
Fast-path implementation for small, focused changes.
Parse the fix description from $ARGUMENTS. If empty, ask: "What needs to be fixed?"
Use Agent(subagent_type: "Explore") to analyze the relevant code and understand the scope of the fix.
Create a single bead for the fix:
br create --title "Fix: <description>" --description "<full context>" --type bug
Track it with TaskCreate(subject: "Fix: <description>", status: "in_progress").
Bootstrap Agent Mail: call macro_start_session(human_key: cwd, program: "claude-code", model: your-model, task_description: "Fix: <description>").
Create a team: TeamCreate(team_name: "fix-<slug>").
Spawn a focused implementation agent with run_in_background: true:
Agent(
subagent_type: "general-purpose",
isolation: "worktree",
name: "fix-impl",
team_name: "fix-<slug>",
run_in_background: true,
prompt: "
## Agent Mail Bootstrap
Call macro_start_session(human_key: '<cwd>', program: 'claude-code', model: 'claude-sonnet-4-6', task_description: 'Fix: <description>')
Reserve files you will edit: call file_reservation_paths(...) before making changes.
Release reservations when done: call release_file_reservations(...).
Send a completion message to <your-name> via send_message when done.
## Fix
Apply this fix: <description>
Context from codebase analysis:
<analysis>
Keep changes minimal and targeted. Do not refactor unrelated code.
"
)
Save the returned task ID for potential TaskStop use.
If the agent goes idle without reporting, nudge: SendMessage(to: "fix-impl", message: "Please report your status.").
After completion, shutdown: SendMessage(to: "fix-impl", message: {"type": "shutdown_request", "reason": "Fix complete."}).
Update task: TaskUpdate(taskId: "<id>", status: "completed").
Call flywheel_review with action: "hit-me" to get fresh-eyes review.
Show the results and ask: "Looks good to commit?" If yes, mark the bead closed: br update <id> --status closed.