Discover (Stage 1) - Lisa

Scan the current project and generate Gastown-compatible semantic memory.

When to Use

Use this command when:

• Starting migration of an existing project
• Need to understand a codebase's tech stack
• Want to document project constraints and goals
• Preparing for roadmap generation

What This Does

1. Scans package files - Detects runtime, framework, dependencies
2. Reads configuration - Identifies database, auth, deployment targets
3. Analyzes documentation - Extracts constraints, personas, non-goals
4. Generates memory - Outputs .gt/memory/semantic.json

Execution

1. Follow the instructions in skills/discover/SKILL.md
2. Create .gt/memory/ directory if it doesn't exist
3. Write semantic.json with discovered facts
4. Record all analyzed files in evidence.files_analyzed

Output

Creates/updates .gt/memory/semantic.json with:

• Project identity (name, type, language)
• Tech stack (runtime, framework, database, auth, etc.)
• User personas if documented
• Constraints and non-goals
• Evidence of what was analyzed

Quality Gates

• semantic.json is valid JSON
• project.name is populated
• At least 2 tech_stack fields are populated
• evidence.files_analyzed has at least 1 entry

Example

/lisa:discover

After running, check .gt/memory/semantic.json for results.

Migration Note

This command replaces the deprecated /lisa:analyze command. The functionality is identical, but the naming better reflects the staged architecture.