Menu
Analyze email, calendar, and file patterns to discover repeatable workflows that AI agents can automate.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Stand up a secure, empty BigQuery data lake on GCP — layered datasets, least-privilege service accounts, Workload Identity (no downloadable keys), read-only authorized views, audit logging, and a budget alert. Agent-guided or Terraform. You connect your own data sources afterward — on purpose.
Set up and configure cloud secrets management for your AI agent. Guides users through platform selection (GCP, AWS, Azure, 1Password, Doppler, HashiCorp Vault), account setup, CLI configuration, and secure credential storage.
Audit and maintain a data semantic layer for AI agents. Scans BigQuery datasets, GCP APIs, secrets, and service accounts to keep your data catalog current.
Email management for small businesses. Daily triage, vendor communication, customer replies, templates, and inbox automation. Built on the gws CLI.
Slack team messaging for small businesses. Send messages, manage channels, post automated updates, run daily standups, and coordinate your team. Includes SMB channel structure and team communication best practices.
Create, query, audit, and optimize Google Ads campaigns via the Google Ads API (Python SDK). Use for campaign management, performance analysis, keyword optimization, and ad creation.
Based on SOC occupation classification
| name | playbook-discovery |
| version | 1.0.0 |
| tier | automation |
| description | Analyze email, calendar, and file patterns to discover repeatable workflows that AI agents can automate. |
| requires | {"bins":["gws"],"skills":["google-workspace","gmail"]} |
Discover repeatable workflows from your historical data that AI agents can automate.
This skill analyzes your business communication data (email, calendar, files, chat) to identify "playbooks" — documented, repeatable workflows with clear triggers, steps, and end states.
Why this matters: Before you can automate, you need to know what to automate. Most small business owners have dozens of repeatable workflows buried in their daily habits — they just haven't documented them. This skill surfaces those patterns.
At least one of these data sources connected:
gmail skill) or Microsoft 365 (Graph API)google-calendar skill) or OutlookMore data sources = better pattern recognition.
Collect data systematically to avoid API limits. Chunk by time period (monthly).
For each data source, extract:
For each of the last 6 months:
Full 6-month period:
Analyze collected data for these pattern types:
People Patterns
Topic Patterns
Temporal Patterns
Flow Patterns
Group related patterns into candidate workflows. For each candidate, define:
| Field | Description |
|---|---|
| Trigger | What kicks off this workflow? (email type, calendar event, time of year, etc.) |
| Steps | What actions happen in sequence? |
| Inputs | What data/information is needed? |
| Outputs | What gets produced? |
| End State | What does "done" look like? |
| Edge Cases | What can go wrong? When should it escalate to human? |
Rank candidate workflows by:
Present findings in this structure:
Emails analyzed: X,XXX (inbox: X,XXX, sent: X,XXX)
Calendar events: XXX
Files reviewed: XXX
Time period: [start] to [end]
Major themes across people, topics, time, and flows.
For each playbook:
## Playbook: [Name]
**Evidence:** What data supports this pattern?
**Trigger Conditions:**
- [Specific trigger 1]
- [Specific trigger 2]
**Step-by-Step Workflow:**
1. [Step 1]
2. [Step 2]
3. [Step 3]
...
**Inputs Required:**
- [Input 1]
- [Input 2]
**Outputs Produced:**
- [Output 1]
- [Output 2]
**Success Criteria:**
- [What does "done" look like?]
**Edge Cases & Escalation:**
- [When to escalate to human]
- [What can go wrong]
**Business Impact:**
[Why this matters — time saved, errors prevented, etc.]
| Playbook | Trigger | Frequency | Impact | Automation Potential |
|---|---|---|---|---|
| [Name] | [Trigger] | Daily/Weekly/Monthly | High/Med/Low | High/Med/Low |
After presenting playbooks, offer:
Playbook: Weekly Partner Status Report
Evidence: 47 emails with subject containing "weekly update" or "status report" sent every Monday between 9-11am to the same 5 recipients over 6 months.
Trigger: Monday 9am OR partner requests update
Steps:
- Pull metrics from dashboard
- Summarize key wins/blockers
- Draft email with standard template
- Send to partner distribution list
Automation Potential: HIGH — template-based, data-driven, predictable schedule
gmail — Email data collectiongoogle-calendar — Calendar data collectionslack-directory — Communication pattern analysisrevenue-forecaster — If a discovered playbook drives or depends on revenue, feed the forecaster's output into itautoresearch — Once a playbook is running, use autoresearch to iteratively improve its outputThe best automation starts with understanding what you already do.