[Fleet Execution] Implement Tier 2 Workflow Agent Skills

[jules-fleet]

Goal Reference

This issue was generated from .fleet/goals/sdk-examples.md. The worker agent should read this file for full context including verification commands, constraints, and structural guidance.

Objective

Implement Tier 2 Agent Skills for design workflows specifically targeting design iteration and design system extraction, which teach agents how to evaluate outputs, iterate, and reconcile tokens.

Code-Level Diagnosis

Code path: packages/sdk/examples/
Mechanism: Multi-step, evaluative design workflows are absent from the current examples directory, omitting crucial guidance on how agents should interactively refine designs or merge token sets.
Root cause: Agent skills representing iterative design flows and system extraction have not been implemented.

Current Implementation

// No workflow-focused examples or extraction scripts for token reconciliation

Proposed Implementation

Files to modify: Scaffold directories for design-iteration and design-system-extraction. Each must include a SKILL.md outlining agent workflow logic, any supporting scripts (like extract-tokens.ts), and a README.md. Update SDK README with references.

Integration (Before -> After)

+ // Workflow skills and extraction helpers added to empower iterative agent execution

Test Scenarios

1. Extraction script validation: The extract-tokens.ts helper script can parse and merge multiple Tailwind configuration blocks successfully.
2. Structure check: SKILL.md documents explicit agent interaction steps and decision-making logic.

Target Files

• packages/sdk/examples/design-iteration/SKILL.md
• packages/sdk/examples/design-iteration/README.md
• packages/sdk/examples/design-system-extraction/SKILL.md
• packages/sdk/examples/design-system-extraction/scripts/extract-tokens.ts
• packages/sdk/examples/design-system-extraction/README.md
• packages/sdk/README.md

Boundary Rules

Restrict your modifications exclusively to the files listed in the Target Files section. Ensure your source changes are entirely backward-compatible if unowned tests outside your boundary fail. Retain all existing file names and locations outside your explicitly declared target list.

---

Fleet Context

• Source: jules:session:13642316243044202608
• Link: https://jules.google.com/session/13642316243044202608