菜单
Automated sprint planning and execution from technical specifications (prompt generation, dependency planning, stateful execution)
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
Check AI CLI usage/quota for Claude Code, OpenAI Codex, Google Gemini CLI, Z.AI, and Synthetic. Use when user asks about remaining quota, usage limits, rate limits, or wants to check how much capacity is left.
Detect installed AI coding CLIs and local model providers; outputs a cached JSON inventory for routing (/detect-clis).
Execute prompts from ./prompts/ directory with various AI models. Use when user asks to run a prompt, execute a task, delegate work to an AI model, run prompts in worktrees/tmux, or run prompts with verification loops.
Orchestrate multi-prompt execution with phase groups, optional auto-deps, and model-tiered agent roles.
Create agent-team orchestrated prompt bundles (orchestrator + sub-prompts) and store them through prompt-manager.
Read and manage daplug configuration from CLAUDE.md using <daplug_config> blocks, with legacy fallback and migration support.
| name | sprint |
| description | Automated sprint planning and execution from technical specifications (prompt generation, dependency planning, stateful execution) |
| allowed-tools | ["Bash(git:*)","Bash(jq:*)","Bash(npx cclimits:*)","Bash(python3:*)","Bash(find:*)","Bash(ls:*)","Bash(cat:*)","Bash(mkdir:*)","Bash(realpath:*)","Read","Write","Edit","Glob","Grep"] |
This skill turns a technical specification into:
/run-prompt)sprint-plan.md).sprint-state.json) for long-running execution# From a spec (generates new prompts)
python3 skills/sprint/scripts/sprint.py <spec-file-or-text> [options]
# From existing prompts (no spec argument)
python3 skills/sprint/scripts/sprint.py --from-existing [options]
Options:
--output-dir DIR (default: ./prompts/) Where to write generated prompt files--plan-file FILE (default: ./sprint-plan.md) Where to write the plan markdown--dry-run Generate plan without creating prompt files or state--from-existing Analyze existing prompts in --output-dir instead of generating from a spec--prompts LIST Include only specific prompts (e.g., 001-005,010)--folder PATH Only include prompts from this subfolder of --output-dir (e.g., providers/)--exclude LIST Exclude specific prompts (e.g., 003,007)--auto-execute Execute phases immediately, updating .sprint-state.json--models LIST Comma-separated models (default: claude,codex,gemini)--max-parallel N Max concurrent prompts per phase (default: 5)--worktree Use worktree isolation when auto-executing--loop Use verification loops when auto-executing--max-iterations N (default: 3) Verification loop max iterations--completion-marker TEXT (default: VERIFICATION_COMPLETE) Loop completion marker--state-file FILE (default: .sprint-state.json) State file path--json Print JSON output (plan + state summary)python3 skills/sprint/scripts/sprint.py status
python3 skills/sprint/scripts/sprint.py add "Implement caching layer"
python3 skills/sprint/scripts/sprint.py remove 005
python3 skills/sprint/scripts/sprint.py replan
python3 skills/sprint/scripts/sprint.py pause
python3 skills/sprint/scripts/sprint.py resume
python3 skills/sprint/scripts/sprint.py cancel --yes
python3 skills/sprint/scripts/sprint.py history
Notes:
.sprint-state.json in the current directory unless --state-file is provided.cancel is destructive for sprint-created worktrees (it removes worktree directories/branches recorded in state).# Analyze all prompts in prompts/
python3 skills/sprint/scripts/sprint.py --from-existing --dry-run
# Only specific prompts
python3 skills/sprint/scripts/sprint.py --from-existing --prompts 001-005,010 --dry-run
# Only prompts in a subfolder
python3 skills/sprint/scripts/sprint.py --from-existing --folder providers/ --dry-run
# Exclude certain prompts
python3 skills/sprint/scripts/sprint.py --from-existing --exclude 003,007 --dry-run
# Combine with execution options
python3 skills/sprint/scripts/sprint.py --from-existing --worktree --loop --auto-execute