A recurring automated workflow that runs on a defined schedule — monitoring rankings, drafting content, generating reports, or tracking competitors. 15 templates across four categories: monitoring, reporting, engagement, and prospecting.
Live example

App walkthrough

How a request flows through the app, from prompt to deliverable.

Cellino Plumbing, Heating, Cooling & Electric
ARP 2.0
SoLV 84%
KW hvac buffalo
Run a geogrid scan for 'hvac buffalo' on Cellino and show me where I rank
Running a 25-point geogrid scan now. Here are your results:
Geogrid Scan ARP 2.0 · 84% Top 3 New
Ask Claude anything about local SEO...
Geogrid Scan
"hvac buffalo"
Cellino Plumbing, Heating, Cooling & Electric
2.0
ARP
84%
SoLV
1.1
ATRP
3
2
1
2
4
2
1
1
2
3
2
1
1
4
5
3
1
3
2
6
5
4
3
5
7
1-3
4-7
8-10

Scheduled Task

A scheduled task is a recurring automated workflow defined by a TASK.md template. It specifies which skills to invoke, what data to pull, when to run, where to write output, and what level of human approval is required.

Definition

Local SEO Skills includes 15 task templates organized into four categories:

  • Monitoring (M-series) — watch for changes: rankings, reviews, GBP edits, LSA positions, AI visibility
  • Reporting (R-series) — produce deliverables: weekly reports, monthly client reports, multi-location rollups, QBRs
  • Engagement (E-series) — create content: GBP post drafts, review response drafts, citation audits, page content audits
  • Prospecting (P-series) — support sales: prospect audits, competitor monitoring

Each task runs on a schedule (daily, weekly, monthly, quarterly, or on-demand) and operates under one of three approval tiers: autonomous, queue, or notify.

Task Components

A TASK.md file contains:

  • Metadata — name, description, schedule, approval tier, required skills and MCPs
  • Fallback guidance — what to do if a tool or data source is unavailable
  • Verification checklist — prerequisites the agent checks before execution
  • Prompt template — execution instructions with variable placeholders
  • Output configuration — file paths and naming conventions
  • Alert thresholds — conditions that trigger notifications

How Tasks Connect to Briefs

Every task writes its output into the client’s brief directory structure. The brief’s session log gets a one-line entry. Findings get updated if something critical surfaces. The output follows a standardized schema for consistent formatting.

  • Brief — the persistent state container that tasks write to
  • Approval Tier — autonomous, queue, or notify — controls human oversight
  • Output Schema — the standardized format for task output files