blog: Goose became its own Watchdog

[iandouglas]

work in progress

[blackgirlbytes]

would be cool to link your scanner here!

[blackgirlbytes]

okay i know we agreed on the title but re reading this..I'm almost thinking a more straightforward title will give it more hits on SEO..or adding something straightforward 🤔 oh and something that puts us at the forefront as experts..How we're keep AI workflows secure or something..that's a little boring..but let me thik on it.

[angiejones]

Suggested change

	That's exactly the challenge we face with Goose recipes. We're building a community cookbook where you can try Goose recipes from other users with confidence that they're safe. (More on that to come!!) But we needed a way to make sure every recipe was safe to run.
	That's exactly the challenge we face with Goose recipes. We're building a community cookbook where you can try Goose recipes from other users with confidence that they're safe. But we needed a way to make sure every recipe was safe to run.

[angiejones]

Suggested change

	I gave it a lot more direction: I wanted the scanner to run from GitHub Actions, and I wanted the scanning to be done within a Docker container so we didn't have to worry about accessing anything sensitive.
	I gave it a lot more direction: I wanted the scanner to run from GitHub Actions, and I wanted the scanning to be done within a Docker container so it could inspect the recipe in an isolated environment.

[angiejones]

Suggested change

	When someone submits a recipe via our GitHub issue template, it will start an automated scan. Goose spins up in an isolated Docker container, analyzes the recipe using its own security expertise plus some of our guidance, scores it for risk, and posts the results right back to the GitHub issue. If it gets approved, it moves onto the next stage. More on that another time!
	When someone submits a recipe via our GitHub issue template, it will start an automated scan. Goose spins up in an isolated Docker container, analyzes the recipe using its own security expertise plus some of our guidance, scores it for risk, and posts the results right back to the GitHub issue.

[angiejones]

for "tutorials and videos", it would be nice to hyperlink these

[angiejones]

Suggested change

	Then we go into "headless" mode. We
	Then we go into "headless" mode:

[angiejones]

Suggested change

	Inside the container, we're installing Goose, and passing in a configuration file for which AI provider and model we want to use, plus a "base" recipe that tells Goose how to analyze the user's recipe. That recipe is also reinforcing a role of Goose being a security expert, but we had to add a little more.
	Inside the container, we're installing Goose, and passing in a configuration file for which AI provider and model we want to use, plus a "base" recipe that tells Goose how to analyze the user's recipe. That recipe is also reinforcing a role of Goose being a security expert.

[angiejones]

this should go up earlier when you are talking about this topic.

[angiejones]

Suggested change

	When our community recipes launches, anyone can submit a recipe knowing it'll get a fair, thorough review. And when you see a recipe get a security approval, remember: it was approved by Goose itself.
	Anyone can submit a recipe knowing it'll get a fair, thorough review. And when you see a recipe get a security approval, remember: it was approved by Goose itself.

[angiejones]

Suggested change

	- iand
	- ian

this is breaking the build

[angiejones]

need the static url

[angiejones]

need the static url

[angiejones]

couple of small fixes needed