Skip to content
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
24 commits
Select commit Hold shift + click to select a range
9bdf302
Add Copilot on Rails evaluation system
alexweininger Aug 11, 2026
eaf0cd5
Merge remote-tracking branch 'origin/feat/CoR' into copilot-on-rails-…
alexweininger Aug 11, 2026
1c4b7e0
Add per-run evaluation diagnostics
alexweininger Aug 12, 2026
4c728fd
Continue evaluation after quality failures
alexweininger Aug 12, 2026
fa4d035
eval: per-stage repair budgets, gate progress metrics, accessibility …
alexweininger Aug 13, 2026
752327a
eval: split the browser gate into a load contract and an advisory jou…
alexweininger Aug 13, 2026
9a475b4
eval: make the debugger gate reachable for Node and browser debug con…
alexweininger Aug 13, 2026
e6c999e
eval: remove artifacts written by a mis-rooted eval launch
alexweininger Aug 13, 2026
d50931f
eval: add two-tier deployment evidence and make gate coverage compile…
alexweininger Aug 13, 2026
65ffe24
eval: detect stalled agent turns and widen gate applicability
alexweininger Aug 14, 2026
497f65d
eval: add security gate and external deploy-skill injection
alexweininger Aug 14, 2026
67a7edd
eval: give the 7 auth scenarios executable security contracts
alexweininger Aug 14, 2026
9efd2d4
eval: stop counting an unevaluated deploy gate as a pass
alexweininger Aug 14, 2026
d6e5674
eval: stop blaming the product for blocked container-image pulls
alexweininger Aug 14, 2026
2fd7244
eval: refuse to run when the evaluator source has drifted
alexweininger Aug 14, 2026
e1119c0
eval: make the stall threshold configurable for higher worker counts
alexweininger Aug 14, 2026
bce4612
Certify gates against the golden fixture and fix a no-op debugger gate
alexweininger Aug 14, 2026
ac1238c
Certify the security and deployment gates against reference projects
alexweininger Aug 14, 2026
9c11bad
Widen scenario contracts so certified gates actually run
alexweininger Aug 14, 2026
6b1c0a3
fix(evals): repair Azurite key that failed the worker gate 100% of th…
alexweininger Aug 15, 2026
a0d722e
fix(evals): stop counting ACA infrastructure errors as grader failures
alexweininger Aug 15, 2026
3e233a0
fix(evals): treat host-toolchain errors as infrastructure, not grader…
alexweininger Aug 15, 2026
3416d4c
chore(evals): add eval:cor:gates:health script
alexweininger Aug 15, 2026
8d2266e
fix(evals): stop charging the product for cascade blocks and outages
alexweininger Aug 17, 2026
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
396 changes: 396 additions & 0 deletions .github/workflows/copilot-on-rails-evals.yml

Large diffs are not rendered by default.

1 change: 1 addition & 0 deletions .gitignore
Original file line number Diff line number Diff line change
Expand Up @@ -67,3 +67,4 @@ testWorkspace
test-results.xml
dist
stats.json
evals/results/
29 changes: 29 additions & 0 deletions .vally.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,29 @@
paths:
skills: []
evals:
- evals/vally
results: evals/results/vally-native
evalFilenames:
- eval.yaml

suites:
copilot-on-rails-offline:
description: Offline contracts for archived Copilot on Rails summary evidence.
evals:
- evals/vally/eval.yaml
copilot-on-rails-hard-gates:
description: Authoritative product and scenario-specific gate contracts.
filter:
evidence: summary-only
copilot-on-rails-native-authoritative:
description: Vally-native ACA execution with fail-closed authoritative release grading.
evals:
- evals/vally/native/authoritative.eval.yaml
copilot-on-rails-native-qualitative:
description: Supplemental and uncalibrated trajectory-plus-diff user-satisfaction grading.
evals:
- evals/vally/native/qualitative.eval.yaml
copilot-on-rails-native-oracle:
description: Offline custom-metrics oracle contract with no live model execution.
evals:
- evals/vally/native/fixtures/oracle.eval.yaml
1 change: 1 addition & 0 deletions .vscodeignore
Original file line number Diff line number Diff line change
Expand Up @@ -12,6 +12,7 @@
build/**
dist/test/**
docs/**
evals/**
gulp*
node_modules/**
out/**
Expand Down
4 changes: 4 additions & 0 deletions eslint.config.mjs
Original file line number Diff line number Diff line change
Expand Up @@ -12,6 +12,10 @@ export default defineConfig([
ignores: [
'api/dist/**',
'api/out/**',
'evals/results/**',
'evals/grader-certification/reference-node-fullstack/**',
'evals/grader-certification/reference-deployable/**',
'evals/vscode-parity/**',
'src/webviews/copilotOnRails/views/react-shim.js',
],
},
Expand Down
Loading
Loading