docs(evals): sirens-echo grades through aos-eval now, so say so #307

Merged

coilysiren merged 1 commit from aos/claude/eval-current-state into main

2026-08-18 17:09:46 +00:00

coilyco-ops commented

2026-08-18 16:58:59 +00:00

Member

One line of truth maintenance after #306 merged.

The page hedged that sirens-echo "shares the boundary declaration shape today, not the dataset", because that was accurate this morning. It is not any more: coilyco-gaming/sirens-echo#957 changed board-deep to emit aos-eval's dataset: key with each record carrying Sample plus its output, added eval/aos-eval-profile.yaml, and pinned the contract with a test on that side. aos-eval validate reads an emitted dataset and reports "10 samples match the profile".

So the pairing rule has one implementation across the two runners rather than two, and this page now says that instead of hedging it.

pre-commit run --all-files passes.

🤖 Generated with Claude Code

One line of truth maintenance after #306 merged. The page hedged that sirens-echo "shares the boundary declaration shape today, not the dataset", because that was accurate this morning. It is not any more: coilyco-gaming/sirens-echo#957 changed `board-deep` to emit aos-eval's `dataset:` key with each record carrying `Sample` plus its `output`, added `eval/aos-eval-profile.yaml`, and pinned the contract with a test on that side. `aos-eval validate` reads an emitted dataset and reports "10 samples match the profile". So the pairing rule has one implementation across the two runners rather than two, and this page now says that instead of hedging it. `pre-commit run --all-files` passes. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

coilyco-ops added 1 commit

2026-08-18 16:59:00 +00:00

docs(evals): sirens-echo grades through aos-eval now, so say so

ci / test (pull_request) Successful in 52s

Details

d6fde8e88a

The page hedged because the claim was forward-looking. It is not any more:
that repo's board-deep emits aos-eval's dataset shape, its profile declares its
one column, and a contract test pins both sides. The pairing rule has one
implementation across the two runners rather than two.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-authored-by: Kai Siren <coilysiren@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>

coilyco-ops requested review from coilysiren

2026-08-18 16:59:00 +00:00