@discovery
Feature: Shell Asset Extraction
  As a developer migrating from Plandex to CleverAgents
  I need to extract and catalog shell scripts from the Go codebase
  So that I can create Python replacements and understand dependencies

  Background:
    Given the Plandex Go repository exists
    And the ShellAssetExtractor is initialized

  Scenario: Extract shell scripts from repository
    When I run the shell asset extraction
    Then shell scripts should be discovered
    And each script should have metadata extracted

  Scenario: Analyze start_local.sh script
    When I extract shell assets
    Then the start_local.sh script should be found
    And it should be marked as requiring Docker
    And it should be marked as requiring Git
    And it should be suggested for Python replacement

  Scenario: Identify environment variables in scripts
    When I extract shell assets
    Then environment variables should be identified
    And each variable should be mapped to its scripts

  Scenario: Detect script dependencies
    When I extract shell assets
    Then scripts requiring Docker should be identified
    And scripts requiring Git should be identified
    And command dependencies should be captured

  Scenario: Generate Python wrappers for scripts
    When I extract shell assets
    And I save the catalog with wrappers
    Then Python wrapper files should be created
    And wrappers should include documentation
    And wrappers should preserve environment variable handling

  Scenario: Save shell asset catalog
    When I extract shell assets
    And I save the catalog
    Then a YAML catalog file should be created
    And a JSON catalog file should be created
    And both catalog files should contain the same data

  Scenario: Identify idempotent scripts
    When I extract shell assets
    Then idempotent scripts should be marked
    And scripts with mkdir -p should be idempotent
    And scripts with test checks should be idempotent

  Scenario: Categorize scripts by purpose
    When I extract shell assets
    Then test scripts should be identified
    And development scripts should be identified
    And code generation scripts should be identified