synthetic-walkthrough-run

Orchestrate synthetic user walkthroughs for multiple personas and generate scored HTML decks.

1|2|Updated Apr 1, 2026
One-click install
npx skills add https://github.com/dimagi-internal/ace --skill synthetic-walkthrough-run
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: synthetic-walkthrough-run
Source: https://github.com/dimagi-internal/ace/tree/main/skills/synthetic-walkthrough-run
Command: npx skills add https://github.com/dimagi-internal/ace --skill synthetic-walkthrough-run

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill eliminates the manual, repetitive work of running synthetic walkthroughs for multiple product personas during ACE Phase 7, removing the need to manually dispatch walkthrough tools, collect screenshots, and aggregate evaluation scores for each persona.

Core Features & Use Cases

  • Per-Persona Walkthrough Orchestration: Automatically dispatches the upstream canopy:walkthrough skill for each defined persona spec, handling all browser automation, scene capture, and score aggregation internally.
  • Timestamped Output Management: Generates uniquely timestamped folders for each persona's walkthrough, storing HTML slideshow decks, per-scene screenshots, and evaluation JSON files in the designated run folder.
  • Append-Only Run State Tracking: Automatically updates the current ACE run's state file to record walkthrough artifacts and scores, maintaining a per-run history of walkthroughs without chaining data across separate runs. Use case: For a Connect opportunity with 4 defined user personas, run this skill to automatically generate full scored walkthrough decks for all personas in one execution, instead of running each walkthrough manually and compiling results by hand.

Quick Start

Use the synthetic-walkthrough-run skill to generate scored walkthrough slideshows for all personas defined in the current ACE opportunity's Phase 7 synthetic specs.

Frequently Asked Questions about synthetic-walkthrough-run

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate persona walkthroughs for multiple users instead of running them manually?

Automated persona walkthrough orchestration dispatches the upstream canopy walkthrough skill for each defined persona spec, handling browser automation, scene capture, and score aggregation internally. This eliminates manual dispatch and compilation for multi-persona product testing.

What outputs are generated when running synthetic user walkthroughs?

Synthetic walkthroughs generate timestamped output folders containing HTML slideshow decks, per-scene screenshots, and evaluation JSON files. These artifacts are staged in designated run folders for each persona walkthrough.

How does scored walkthrough evaluation work for product personas?

Scored walkthrough evaluation uses LLM-as-Judge metrics aggregated for each defined user persona tied to a Connect opportunity. The system automatically collects and compiles these evaluation scores alongside the generated HTML walkthrough decks.

Do I need to validate plugins and credentials before running synthetic walkthrough orchestration?

Yes, pre-flight validation of required plugins, binaries, and credentials is performed automatically before synthetic walkthrough orchestration. This ensures the environment meets all prerequisites before dispatching browser automation tasks.

How is run state managed across multiple synthetic walkthrough executions?

Run state management uses append-only updating of per-run ACE state files to record walkthrough artifacts and scores. This maintains a per-run history of walkthroughs without chaining data across separate executions.

Can I run scored walkthroughs for all personas in a single execution during ACE Phase 7?

Yes, synthetic walkthrough orchestration generates full scored walkthrough decks for all personas defined in an ACE opportunity's Phase 7 synthetic specs in one execution, replacing manual per-persona runs and hand compilation.