Specification-driven development workflow for AI agents. Use when tasks are too large for a single session, require multi-step implementation, span multiple files/features, or need persistent requirements tracking.
A CLI tool for managing structured specifications that persist across sessions, with JSON output designed for AI agents.
The CLI is located at ./scripts/specdev-cli/ relative to this SKILL.md file.
| Platform | Script |
|---|---|
| Unix/Linux/macOS | specdev |
| Windows | specdev.ps1 |
Claude Code: Use ${CLAUDE_PLUGIN_ROOT}/skills/specdev/scripts/specdev-cli/specdev (or specdev.ps1 on Windows).
| Command | Description |
|---|---|
init | Initialize .specs/ structure (idempotent) |
new <name> | Create new spec with templates |
list | List all active specs and progress |
context <spec> | Get spec context (--level min|standard|full) |
path <spec> | Get spec directory path |
archive <spec> | Move completed spec to archived |
validate <path> | Check spec file completeness |
hook <event> | Run hook for event |
Use when starting a new feature or complex task.
specdev init (if no .specs/ exists)specdev new <name> to create spec structure with templatesspec.md, plan.md, and tasks.yaml based on initial understandingAskUserQuestion to clarify ambiguities or validate assumptionsspecdev validate <name> to verify spec structureTip: Always include verification methods in tasks (tests, assertions, manual checks). This creates feedback loops that catch errors early. Ask user...