automation

name	automation
description	Trigger-driven automation definitions and runner management. Use when creating, running, or managing skill-based automation pipelines.
chains_with	["quality","automated-learning"]

Automation Skill — Always-On Skill Pipelines

Overview

Automations connect triggers (events, schedules, file changes) to skill pipelines (our 25 skills) to produce actions (PRs, comments, Slack messages, reports).

Each automation is a YAML file in automations/definitions/ that declares:

• A trigger (GitHub event, cron schedule, file change, CLI)
• A skill pipeline (ordered list of skills to invoke)
• Actions (what to do with results)

Trigger Types

Type	How to Fire	Example
github	PR opened/pushed	automate.py run pr-security-review
cron	Scheduled via system crontab	automate.py cron-install
file_watch	File changes on disk	automate.py watch .
cli	Manual invocation	automate.py run generate-docs

Commands

# List all available automations
python .opencode/scripts/automate.py list

# Run an automation once
python .opencode/scripts/automate.py run pr-security-review

# Run all automations
python .opencode/scripts/automate.py run-all

# Watch directory for file changes (daemon)
python .opencode/scripts/automate.py watch .

# Install cron jobs from automation schedules
python .opencode/scripts/automate.py cron-install

# View run history
python .opencode/scripts/automate.py status

Adding an Automation

1. Create a YAML file in automations/definitions/
2. Declare trigger, pipeline, and actions
3. Run it: python .opencode/scripts/automate.py run <name>

Pipeline Design

Pipelines compose skills sequentially. Each step can be:

• A skill name string: "security"
• A skill object with config: {skill: "security", config: {severity: "high"}}
• A conditional step: {skill: "security", when: "severity == critical"}

Automation Templates

Template	Trigger	Pipeline	Action
PR Security Review	GitHub PR	exhaustive-crosscheck → security → code-hardener	PR comment + Slack alert
Nightly Audit	Cron	quality → security → performance	File report + log
Add Test Coverage	GitHub PR	testing → quality → python	PR comment
Find Critical Bugs	Cron	exhaustive-crosscheck → debugging → code-hardener	Draft PR + log
Generate Docs	CLI	documentation → communication	Git commit
Weekly Summary	Cron	communication → data-science	File report