evalkit-py 0.2.0.dev12__py3-none-any.whl → 0.2.0.dev13__py3-none-any.whl

{evalkit_py-0.2.0.dev12.dist-info → evalkit_py-0.2.0.dev13.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: evalkit-py
-Version: 0.2.0.dev12
+Version: 0.2.0.dev13
 Summary: Standalone agent-eval product: server, CLI, UI, finance eval suite, and EDD skills
 Author-email: InferEdge Tech <tech@inferedge.com>
 License-File: LICENSE
@@ -50,7 +50,7 @@ in; a better agent comes out, with evals your domain experts signed off on.
 
 ```mermaid
 flowchart LR
-  T["Your agent's<br/>real runs"] -->|mine| C["Draft cases"]
+  T["Your agent's<br/>real runs"] -->|mine| C["New cases"]
   C -->|review| R["Approved cases"]
   R -->|run| S["Scores +<br/>trajectories"]
   S -->|diagnose| F["File-cited fixes"]
@@ -66,7 +66,8 @@ flowchart LR
    one, groups the failure patterns and drafts eval cases.
 2. **Agree on what right looks like.** Domain experts read each case as a
    question and the facts a right answer must contain, then approve it,
-   comment, or send it back with changes. Only approved cases count.
+   comment, or send it back with changes. Approved cases are promoted to
+   your production suite.
 3. **Measure it.** evalkit runs the cases against your agent, keeps the full
    trajectory of every run (messages, reasoning, tool calls, tokens, cost,
    timing) and scores it with LLM-judged rubrics, built-in tool-call checks
@@ -94,11 +95,11 @@ same cases.
   year-end." An expert can read and judge it without reading code. The
   case-authoring skill interviews the expert in their own terms and writes
   the YAML.
-- **Review in the web UI.** Draft → Review requested → Approved, or Changes
-  requested with a comment. Threaded comments sit on the case, and each
-  case keeps a changelog of every sync and decision. Runs on shared
-  datasets use approved cases by default, so a half-written case never
-  moves the score.
+- **Review in the web UI.** Unreviewed → Review requested → Approved, or
+  Changes requested with a comment. Threaded comments sit on the case, and
+  each case keeps a changelog of every sync and decision. Review never
+  filters a run: every case runs, and approval decides which version is
+  promoted.
 - **Workspaces.** Each developer has a personal workspace to iterate in
   without touching anyone's results. Shared workspaces such as `dev` and
   `prod` hold the team's datasets, agents and experiments, with `writer`

{evalkit_py-0.2.0.dev12.dist-info → evalkit_py-0.2.0.dev13.dist-info}/RECORD

@@ -364,8 +364,8 @@ evalkit/bundled/skills/evalkit-deployment/try/run.py,sha256=xS29xgHmK4pDwPrUVqpt
 evalkit/bundled/skills/evalkit-usage/references/suite-fixtures.md,sha256=plz5MvzIG4qQaMUkuPDcqIcf4M1n_Wn_G5zfXZPNctU,1814
 evalkit/bundled/skills/evalkit-deployment/try/cases/acme/balance-sheet/cases.yaml,sha256=IowaNJaLwa-sahiqpvw9xX_Mj3TrfeVZX93f_RfENU0,802
 evalkit/bundled/skills/evalkit-deployment/try/cases/acme/income-statement/cases.yaml,sha256=tf2DalZVlSI4Km-w6B0ER13bLGXEutY0_LxFO0nIroA,1235
-evalkit_py-0.2.0.dev12.dist-info/METADATA,sha256=bDkob244hEqCs5kkxMyR8zGvIQNQD2h4Y4bvocm6q8U,8484
-evalkit_py-0.2.0.dev12.dist-info/WHEEL,sha256=2iZrJkDIkeDh0k9SikKMqFuRP-ua6mb4Ir1AsER4BEw,87
-evalkit_py-0.2.0.dev12.dist-info/entry_points.txt,sha256=wugnobwgFyE6Kotv-Z9R0Xk4Um1ROfx5UU-DntvQQVA,49
-evalkit_py-0.2.0.dev12.dist-info/licenses/LICENSE,sha256=hIahDEOTzuHCU5J2nd07LWwkLW7Hko4UFO__ffsvB-8,34523
-evalkit_py-0.2.0.dev12.dist-info/RECORD,,
+evalkit_py-0.2.0.dev13.dist-info/METADATA,sha256=yecEUTF8Xa8i9_AMBkmGPM1XC1eEzp23mgiTmEAMcGY,8509
+evalkit_py-0.2.0.dev13.dist-info/WHEEL,sha256=2iZrJkDIkeDh0k9SikKMqFuRP-ua6mb4Ir1AsER4BEw,87
+evalkit_py-0.2.0.dev13.dist-info/entry_points.txt,sha256=wugnobwgFyE6Kotv-Z9R0Xk4Um1ROfx5UU-DntvQQVA,49
+evalkit_py-0.2.0.dev13.dist-info/licenses/LICENSE,sha256=hIahDEOTzuHCU5J2nd07LWwkLW7Hko4UFO__ffsvB-8,34523
+evalkit_py-0.2.0.dev13.dist-info/RECORD,,