Run the propose → plan → build → verify lifecycle autonomously across the queue of active changes. Advanced — requires automation config in config.yaml. Use when you want specclaw to advance every change to its next phase without prompting; do not use for a single targeted change (use the specific verb instead).
13 stars
0 votes
0 copies
1 view
Added August 31, 2026
ai-agents
Security analysis
A100/100
Scanned August 31, 2026
$npx -y skills add chan4lk/specclaw --skill auto --agent claude-code
Installs into .claude/skills of the current project.
Are you the author of Auto?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/chan4lk-auto)
---
description: Run the propose → plan → build → verify lifecycle autonomously across the queue of active changes. Advanced — requires automation config in config.yaml. Use when you want specclaw to advance every change to its next phase without prompting; do not use for a single targeted change (use the specific verb instead).
---
# specclaw auto
**First, run** `specclaw-ensure-init .specclaw` — idempotently creates `.specclaw/` if it doesn't exist (silent if already initialized; auto-inits using the current directory's basename as the project name).
Run the lifecycle autonomously for the active queue.
1. Read `.specclaw/STATUS.md` for the next actionable item across changes.
2. For each change, advance it to its next phase:
- Proposal exists without plan → run `/specclaw:plan`.
- Plan exists without implementation → run `/specclaw:build`.
- Built without verification → run `/specclaw:verify`.
3. Respect `config.yaml` limits:
- `automation.max_tasks_per_run` — cap on total tasks executed this run.
- `automation.auto_verify` — whether to chain verify after build.
- `automation.auto_archive` — whether to chain archive after a passing verify.
4. Notify the user of results via the configured notification channel.
Stop and surface the issue if any phase fails, rather than blindly advancing to the next change.
