feat: updated scaffolding template

[zilto]

This updates the template with several changes:

1. more descriptive module-level docstring; removes keyword vibe coding and specifics about dlt commands
2. fix import path from dlt.sources.rest_api.typing. The previous import was unintended and creates linting errors with pylance (this throws off IDE agents like Cursor)
3. Remove instructions to set API key in plaintext. This suggests that the LLM should read / write your API key. Instead, added a comment about how dlt.secrets.value works
4. Remove get_data() function. This scope is unecessary and adds complexity if we intend to run this as a script.
5. Remove dataset_name kwarg. This simplifies the work of the LLM. Anyways, the name is derived from the pipeline name.
6. Use dev_mode=True, this avoids schema migrations across runs.
7. Use refresh="drop_sources" to avoid state conflicts as the agents does multiple iterations

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	docs	b93a710	Commit Preview URL Branch Preview URL	Nov 12 2025, 10:04 PM

[rudolfix]

LGTM, thx for fixing this

[rudolfix]

Note: I've modified AST parser to allow optional dataset names and script import checker to allow pipelines to be instantiated on import (makes sense also for runtime)

pipeline running, iterating resources and source instantiation is not allowed

[zilto]

Converted to draft to avoid merging. I realized that dev_mode has some counterintuitive behavior #3276 which may be confusing for humans / LLMs.

I think the workspace feature and automated local pipelines_dir is generally helpful for LLMs since it gives the full "project context/state" under the IDE project root.