菜单
Simplified state file management with consolidated FLOW.md (project) and ITEM-XXX.md (per-item) files
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
Interactively elicit, capture, and maintain software/system requirements as a living, traceable repository. Use whenever the user wants to gather, write, refine, or organize requirements — including producing or updating an SRS (software/system requirements specification), writing ADRs / architecture decision records, building a requirements traceability graph (DAG), or keeping a decision ledger. Trigger proactively when the working folder already holds requirements artifacts (a `requirements/` folder of YAML, `decisions/` ADRs, an `srs.md`, or `ledger.md`), or when the user shares documents that are clearly specs, RFCs, or feature briefs and wants them structured. Also use for phrases like "let's spec this out", "interview me about what we're building", "capture these requirements", "write a decision record", or "turn these notes into a spec". Not for project management, code review, application code, dependency manifests like requirements.txt, or hardware "system requirements" (RAM/CPU).
Use this skill whenever the user wants to create a bash script that drives an autonomous coding loop — scripts that run Claude Code on a schedule (cron, GitHub Actions, systemd timer) or on-demand to make repository improvements, execute plans, triage issues, or otherwise do agentic work with minimal human intervention. Triggers include phrases like "scout script", "autonomous script", "agent harness", "script that runs on a schedule", "script that implements a plan", ".scripts/*.sh", or any request for a bash script whose body is "run Claude Code with a prompt and commit the results". Always use this skill before writing such a script — it encodes the interview flow, the adversarial-reviewer pattern, and the locking/ledger conventions.
Task structure and atomic commit patterns for granular, verifiable work units
Conflict identification and resolution patterns for requirements, decisions, and plans
Context window management techniques for maintaining efficiency and preventing context bloat
Exploration map management for tracking discussed areas and uncharted territory during DISCUSS phase
| name | state-management |
| description | Simplified state file management with consolidated FLOW.md (project) and ITEM-XXX.md (per-item) files |
| triggers | ["workflow state","state files","FLOW.md","checkpoint","resume","backlog","work item"] |
This skill provides techniques for managing workflow state through simplified, consolidated markdown files. The refactored structure uses just two file patterns: a single FLOW.md at project level and one ITEM-XXX.md per work item.
.flow/
├── FLOW.md # Project state + backlog (single file)
└── items/
├── ITEM-001.md # All state for work item 001
├── ITEM-002.md # All state for work item 002
└── ...
| File | Purpose | Contains |
|---|---|---|
FLOW.md | Project-level state | Vision, backlog, active item, capabilities cache |
ITEM-XXX.md | Per-item state | Phase, decisions, requirements, tasks, checkpoint |
The single project file contains:
See templates.md for the full FLOW.md template.
Each work item has a single consolidated file containing:
See templates.md for the full ITEM-XXX.md template.
Items progress through these phases:
| Phase | Description | Next Phase |
|---|---|---|
BACKLOG | Defined but not started | DISCUSS |
DISCUSS | Gathering requirements | PLAN |
PLAN | Creating task plan | EXECUTE |
EXECUTE | Implementing tasks | VERIFY |
VERIFY | Validating implementation | DONE |
DONE | Completed | (terminal) |
ON_HOLD | Paused intentionally | Previous |
BLOCKED | Waiting on dependency | Previous |
1. Create .flow/ directory if not exists
2. Create .flow/items/ directory
3. Scan for installed plugins
4. Detect project type from file patterns
5. Create FLOW.md with empty backlog + capabilities
/flow-workflow:start)1. Read FLOW.md to get next item number
2. Generate ITEM-XXX ID
3. Create .flow/items/ITEM-XXX.md with DISCUSS phase
4. Update FLOW.md backlog table
5. Set as active item in FLOW.md
1. Update checkpoint in current item's file
2. Update FLOW.md active item pointer
3. Load new item's context
1. Verify completion criteria for current phase
2. Update phase in ITEM-XXX.md
3. Create checkpoint snapshot
4. Initialize next phase section
1. Record current phase and task progress
2. Capture context budget percentage
3. List next recommended action
4. Add timestamp
1. Read FLOW.md to get active item
2. Read ITEM-XXX.md checkpoint section
3. Restore context from checkpoint
4. Identify next action
5. Continue workflow
Track context usage in FLOW.md and report:
## Context Monitor
- Current session: [X]% (limit: 50% before fresh agent)
- Last auto-spawn: [TIMESTAMP]
- Fresh agents this session: [N]
Before any phase transition, validate:
If legacy .flow/ structure detected (BACKLOG.md, ACTIVE.md, PROJECT.md, etc.):
.flow/legacy/State management integrates with:
/flow-workflow:goSee templates.md for full file templates.