Skip to content

MIT-688: eval database names are opaque hashes in decision 0023 - #182

Merged
ZachFara merged 1 commit into
feature/eval-setupfrom
docs/MIT-688
Nov 12, 2026
Merged

ZachFara merged 1 commit into
feature/eval-setupfrom
docs/MIT-688

Conversation

@ZachFara

Copy link
Copy Markdown
Owner

What changed

docs/decisions/0023-behavioral-eval.md gave the old eval database name. I kept the original wording and added an "Updated 2026-11-11 (MIT-653)" note saying sessions now use commons_<16 hex>, a hash of condition, index and scenario, so a store line doesn't reveal the scenario or that it's an eval.

Why

database_name in tools/eval/runner.py changed format in an earlier PR, and the decision record still described the old one. It was the only doc naming the old format.

How tested

grep -rn "commons_eval_" docs/ AGENTS.md DESIGN.md README.md (excluding commons_eval_store*) now only matches the preserved original text. make check passed once.

Learned

Nothing.

🤖 Generated with Claude Code

https://claude.ai/code/session_01Qc1CQTb8r67J32HpkhS43y

@ZachFara
ZachFara merged commit 734a67c into feature/eval-setup Nov 12, 2026
2 checks passed
@ZachFara
ZachFara deleted the docs/MIT-688 branch November 12, 2026 02:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant