685.1-cli-validation Test current prompt CLI and feed back to ls-prompt-ux design

[codekiln]

Parent Issue

Sub-issue of #685 (ls-claude-code-plugin milestone)

Phase

0.1: CLI Validation

Purpose

Before building plugin commands, validate that the current langstar prompt CLI is usable by Claude. Feed findings back to ls-prompt-ux milestone #16 and Issue #679.

Scope

1. Test Current CLI Commands

Run each prompt command and evaluate Claude-friendliness:

• langstar prompt list
• langstar prompt get <handle>
• langstar prompt pull <handle>
• langstar prompt push
• langstar prompt search

2. Evaluate Outputs

For each command:

• Is help text self-explanatory for an AI agent?
• Can Claude parse the output format (table/json)?
• Are error messages actionable?
• Is the command's purpose clear from --help?

3. Document Findings

Create test report documenting:

• Commands that work well for Claude
• Commands that are confusing or have unclear output
• Specific recommendations for 668.0-design Design prompt command structure for AI-first UX and prompt engineering workflows #679 design decisions

4. Feed Back to ls-prompt-ux

Post findings as comment on #679 to inform CLI design decisions.

Deliverables

• CLI test results document
• Comment on 668.0-design Design prompt command structure for AI-first UX and prompt engineering workflows #679 with specific feedback
• Updated scout report if new blockers found

Dependencies

• Requires current langstar prompt commands to be functional
• Informs 668.0-design Design prompt command structure for AI-first UX and prompt engineering workflows #679 design decisions

Success Criteria

• All prompt commands tested and documented
• Feedback incorporated into ls-prompt-ux planning