observal-cli 0.7.0 (new formula)

[Apoorvgarg-creator]

• Have you followed the guidelines for contributing?
• Have you ensured that your commits follow the commit style guide?
• Have you checked that there aren't other open pull requests for the same formula update/change?
• Have you built your formula locally with HOMEBREW_NO_INSTALL_FROM_API=1 brew install --build-from-source observal-cli?
• Is your test running fine brew test observal-cli?
• Does your build pass brew audit --strict observal-cli (after doing HOMEBREW_NO_INSTALL_FROM_API=1 brew install --build-from-source observal-cli)? If this is a new formula, does it pass brew audit --new observal-cli?

---

• AI was used to generate or assist with generating this PR. Please specify below how you used AI to help you, and what steps you have taken to manually verify the changes.

AI disclosure

Claude (Anthropic) was used as a pair-programming assistant to scaffold the formula and the resource blocks. Specifically:

• The runtime dep list (url, sha256, every resource block) was machine-derived from a clean pip install observal-cli==0.7.0 in a throwaway uv venv, then each (url, sha256) pair was fetched directly from https://pypi.org/pypi/<name>/<version>/json. No values were generated from model memory.
• The formula skeleton (depends_on, install method, test block) was adapted by hand from existing homebrew-core Python virtualenv formulae such as octodns.rb and harlequin.rb.
• Manual verification I performed locally on macOS 26.3 (arm64, Tier 2) before opening this PR:
  • brew style observal-cli → no offenses.
  • brew audit --strict --online --new --formula observal-cli → exit 0.
  • HOMEBREW_NO_INSTALL_FROM_API=1 brew install --build-from-source observal-cli → installs from sdist cleanly.
  • brew test observal-cli → exit 0.
  • observal --version from the brew prefix prints observal 0.7.0.
  • ruby -c Formula/o/observal-cli.rb is clean.

---

What is observal-cli?

observal-cli is the command-line interface for Observal, an open-source (AGPL-3.0) self-hosted AI agent registry with built-in observability. The CLI lets users:

• Browse and install AI coding-agent configurations from a registry (think Docker Hub for AI agents).
• Publish their own agents bundled with MCP servers, skills, hooks, prompts and sandboxes.
• Authenticate with a self-hosted Observal server and connect supported IDEs (Claude Code, Kiro, Cursor, Gemini CLI, Codex CLI, OpenCode, VS Code, Copilot CLI) so every interaction produces traces, spans and session telemetry.

Upstream is at https://github.com/BlazeUp-AI/Observal (1,100+ stars, 140+ forks, AGPL-3.0-only, stable Beta — the 0.7.0 sdist published to PyPI carries Classifier: Development Status :: 4 - Beta).

Notes

• Pure-Python virtualenv formula using Language::Python::Virtualenv. The only system deps are rust (build-time, for asyncpg's small native shim when wheels aren't preferred) and libyaml (for pyyaml).
• The optional [migrate] extra (which adds pyarrow for Parquet export/import) is deliberately omitted from the formula; users who need the migration commands can install via pip install 'observal-cli[migrate]' or uv tool install 'observal-cli[migrate]' separately. This avoids dragging Apache Arrow into a default install.
• head is wired to the upstream main branch for HEAD installs.
• Upstream maintains its own tap at https://github.com/BlazeUp-AI/homebrew-observal shipping the PyInstaller release binaries; this PR is the source-build formula intended for homebrew-core.

[botantony]

replace with formula:

depends_on "certifi" => :no_linkage

pypi_packages exclude_packages: "certifi"

[botantony]

Why not the latest Python?

[botantony]

Suggested change

	output = shell_output("#{bin}/observal --version")
	assert_match(/observal/i, output)
	help_output = shell_output("#{bin}/observal --help")
	assert_match "Usage", help_output
	assert_match version.to_s, shell_output("#{bin}/observal --version")

Homebrew requires a simple test that checks if a program can do basic stuff correctly. From Formula Cookbook:

We want tests that don’t require any user input and test the basic functionality of the application. For example foo build-foo input.foo is a good test and (despite their widespread use) foo --version and foo --help are bad tests. However, a bad test is better than no test at all.