Add focused multimodal catalog agent example

[svonava]

Summary

• add a focused catalog agent to the existing taxonomy-classification example
• rank candidate paths from copy and from image plus copy
• ask Qwen3.6-27B to select from the union and return a review flag
• add single-listing and 100-row evaluation commands
• document the 55 initial misses and the measured improvement
• update the example to sie-sdk 0.6.24

Measured result

The current recorded run moves exact-path matches from 42/100 for copy alone to 54/100 for the agent. Macro hierarchical F1 moves from 0.6710 to 0.7446. These values came from real model outputs.

Validation

• uv run ruff format --check .
• uv run ruff check .
• uv run pytest -q
• predict-catalog-agent --help
• eval-catalog-agent --help

Before merge

• replay the full example through one stable SIE Cloud instance
• replace the compact result summary with that Cloud run
• confirm the documented metrics remain unchanged

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: da30895c-9d5b-4de7-8bae-2d7c66d29c3e

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch agent/ecommerce-taxonomy-verifier

---

_{Comment @coderabbitai help to get the list of available commands.}