Feature/evals dod

[reidbaker-agent]

Add proof of concept evals to definition-of-done.

#174

Follow up work coming for a better way to ignore directories because that introduced the pattern of rules based custom arguments.

Pre-launch Checklist

• I read the Contributor Guide and followed the process outlined there for submitting PRs.
• I read the Tree Hygiene wiki page, which explains my responsibilities.
• I read the Flutter Style Guide recently, and have followed its advice.
• I signed the CLA.
• I listed at least one issue that this PR fixes in the description above.
• I updated/added relevant documentation (doc comments with ///).
• I added new tests to check the change I am making, or this PR is test-exempt.
• All existing and new tests are passing.

If you need help, consider asking for advice on the #hackers-new channel on Discord.

[gemini-code-assist]

Code Review

This pull request introduces a rubric-based LLM-as-a-judge evaluation system for agent skills, adding documentation, ignore rules, and a test suite for the definition-of-done skill. Feedback highlights that the new .gitignore rules for evals directories are redundant or ineffective due to Git's directory traversal behavior. Additionally, the evaluation prompt path in evals.json should be updated to use the correct relative path from the repository root to prevent execution errors.

[gemini-code-assist]

In Git, if a parent directory is ignored (due to the * rule at the top), Git will not traverse into it. Therefore, !**/evals/ and !**/evals/** will not work for any third-party/ignored skill folders because their parent directories are excluded.

For the explicitly unignored skill folders (like !definition-of-done/), all of their contents (including evals/) are already unignored by default, making these rules redundant for them.

Since these rules are either redundant or ineffective, they should be removed to keep the .gitignore clean.