Release/v0.10.0

[lchoquel]

[greptile-apps]

Greptile Summary

This PR updates the starter for the hosted Pipelex API. The main changes are:

• Switches runtime dependencies from pipelex to pipelex-sdk and python-dotenv.
• Rewrites the hello-world CLI to send the .mthds bundle through PipelexAPIClient.
• Updates tests, markers, CI commands, and docs for offline checks versus API/inference checks.
• Adds a bootstrap skill and script for renaming the template into a real project.

Confidence Score: 3/5

The update is mostly straightforward, but generated Pipelex runtime artifacts can now be accidentally tracked and the sample API output handling can silently report success without usable text.

The changed dependency and test setup is coherent, but the ignore-rule regression affects common local workflows and the CLI path lacks validation for unexpected API response content.

.gitignore and my_project/hello_world.py

T-Rex Logs

What T-Rex did

• Created representative dummy Pipelex files under .pipelex/storage and .pipelex/traces, verified they are not ignored by Git, and confirmed they can be staged with a targeted git add.
• Migrated from the old local runner path to the new Pipelex API client path; the initial HTTP 200 OK run showed EXTRACT_HELPER_MISSING, and the subsequent run used start_and_wait with pipe_code='hello_world', mthds_contents_count=1, mthds_contents_sha256 equal to BUNDLE_SHA256, and EXTRACT_MAIN and EXTRACT_FALLBACK data.
• Compared bootstrap before and after states; the after state shows dry-run hash unchanged, real bootstrap transformations, and a failure at the missing README clone URL assertion, with the validation script capturing the before/after expectations.
• Reviewed workflow tests and environment setup; the before/after workflow captures indicate validate expansion and exclusions, but pytest collection was attempted and failed early due to ModuleNotFoundError: No module named 'dotenv' before any tests ran.

_{Ran code and verified through T-Rex}

Comments Outside Diff (1)

1. General comment

   README clone instructions are not updated when --clean is used

   • Bug
     • Running the requested bootstrap command with --repo-url https://github.com/acme/invoice-extractor and --clean does not leave README clone instructions pointing at the provided repository. The validation reached grep -q 'git clone https://github.com/acme/invoice-extractor' README.md and exited 1; follow-up grep showed README contains project/package substitutions and the Pipelex API GitHub URL, but no project clone command for the provided repo.
   • Cause
     • In .claude/skills/bootstrap/scripts/bootstrap.py, transform_readme() calls strip_template_block(text) before applying clone replacements. strip_template_block() deletes the ### Use this template subsection up to the next H2, which appears to include the placeholder clone instructions, so lines 262-269 no longer have git clone https://github.com/yourusername/your-repo-name.git / cd your-repo-name text to replace.
   • Fix
     • Apply repository clone replacements before stripping template scaffolding, or adjust strip_template_block() to preserve/move the clone instructions into a real-project setup section when --repo-url is provided. Then add an automated test covering --clean --repo-url together.

   _{Ran code and verified through T-Rex}

Prompt To Fix All With AI

Fix the following 2 code review issues. Work through them one at a time, proposing concise fixes.

---

### Issue 1 of 2
.gitignore:30
**Pipelex Traces Become Trackable**

Self-hosted/local Pipelex runs can still create `.pipelex/storage` and `.pipelex/traces`, but this change stops ignoring those generated artifacts. A normal `git add -A` can now stage runtime traces, payloads, and local state from supported workflows, which can leak sensitive run data into commits.

### Issue 2 of 2
my_project/hello_world.py:66-67
**Missing Text Prints Success**

When the API returns a valid content dictionary without a `text` key, `hello_world()` prints `None` and exits successfully. The e2e test only checks that the coroutine does not raise, so a schema mismatch or unexpected API output can look like a successful run while showing no generated haiku.

```suggestion
    generated_text = content.get("text")
    if not isinstance(generated_text, str) or not generated_text:
        raise RuntimeError("The pipeline returned no text output.")

    print("Your first Pipelex output:\n")
    print(generated_text)
```

_{Reviews (1): Last reviewed commit: "chore: update build-system configuration..." | Re-trigger Greptile}