Agent Learning Kit v1 + unified benchmark harness

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -1,65 +1,35 @@
 name: Bug report
-description: Something is broken in a Python or TypeScript SDK package.
-labels: ["bug", "needs-triage"]
+description: Report a reproducible problem in Agent Learning Kit.
+title: "[Bug]: "
+labels: ["bug"]
 body:
-  - type: markdown
-    attributes:
-      value: |
-        Before filing: check [open issues](https://github.com/future-agi/ai-evaluation/issues) and [Discussions](https://github.com/orgs/future-agi/discussions) — your question may already be answered.
-
-  - type: dropdown
-    id: package
-    attributes:
-      label: Affected package
-      options:
-        - ai-evaluation (Python)
-        - "@future-agi/ai-evaluation (TypeScript)"
-        - Not sure
-    validations:
-      required: true
-
-  - type: input
-    id: version
+  - type: textarea
+    id: summary
     attributes:
-      label: Package version
-      placeholder: "e.g. ai-evaluation==1.1.0 or @future-agi/ai-evaluation@1.0.0"
+      label: Summary
+      description: What failed?
     validations:
       required: true
-
   - type: textarea
-    id: description
+    id: reproduce
     attributes:
-      label: What happened?
-      description: A clear description of the bug. What did you expect, and what did you get instead?
+      label: Reproduction
+      description: Commands, manifest, code, or example needed to reproduce.
+      placeholder: |
+        uv run ...
+        agent-learn ...
     validations:
       required: true
-
   - type: textarea
-    id: repro
+    id: expected
     attributes:
-      label: Minimal reproduction
-      description: The shortest self-contained code that shows the problem. Remove any real API keys.
-      render: python
+      label: Expected Behavior
     validations:
       required: true
-
   - type: textarea
-    id: error
-    attributes:
-      label: Error output / stack trace
-      render: shell
-
-  - type: textarea
-    id: env
+    id: environment
     attributes:
       label: Environment
-      description: Python version, Node version, OS, gateway version or cloud vs self-hosted.
-      placeholder: |
-        Python 3.11.9 / Node 20.12.0
-        macOS 14.4 / Ubuntu 22.04
-        Gateway: Future AGI Cloud (gateway.futureagi.com) / self-hosted v0.9.1
-
-  - type: textarea
-    id: extra
-    attributes:
-      label: Anything else?
+      description: Python, Node, OS, package version, and relevant extras.
+    validations:
+      required: false

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -1,58 +1,41 @@
 name: Feature request
-description: Propose a new metric, guardrail, or capability for ai-evaluation.
-labels: ["enhancement", "needs-triage"]
+description: Propose a new workflow, adapter, evaluator, optimizer, or release gate.
+title: "[Feature]: "
+labels: ["enhancement"]
 body:
-  - type: markdown
-    attributes:
-      value: |
-        For big proposals (new framework integrations, API surface changes), consider opening a [Discussion](https://github.com/orgs/future-agi/discussions) first — it's easier to iterate on design before a PR exists.
-
-  - type: dropdown
-    id: package
-    attributes:
-      label: Which package(s) does this touch?
-      multiple: true
-      options:
-        - ai-evaluation (Python)
-        - "@future-agi/ai-evaluation (TypeScript)"
-        - Not sure
-    validations:
-      required: true
-
   - type: textarea
     id: problem
     attributes:
-      label: What problem does this solve?
-      description: What are you trying to do, and what's getting in your way?
+      label: Problem
+      description: What developer workflow should become easier or more robust?
     validations:
       required: true
-
   - type: textarea
-    id: solution
+    id: proposal
     attributes:
-      label: Proposed solution
-      description: What should the SDK do? A rough API sketch is worth a thousand words.
+      label: Proposal
+      description: Describe the requested API, CLI, example, or gate.
     validations:
       required: true
-
-  - type: textarea
-    id: alternatives
-    attributes:
-      label: Alternatives considered
-      description: What have you tried? Why doesn't it work well enough?
-
   - type: dropdown
-    id: contribution
+    id: surface
     attributes:
-      label: Are you willing to open a PR?
+      label: Surface
       options:
-        - "Yes — I'd like to implement this"
-        - "Maybe — I could help with parts"
-        - "No — hoping someone else picks it up"
+        - simulate
+        - evals
+        - optimize
+        - redteam
+        - suite
+        - TypeScript SDK
+        - release/readiness
+        - documentation
     validations:
       required: true
-
   - type: textarea
-    id: extra
+    id: proof
     attributes:
-      label: Anything else?
+      label: Suggested Proof
+      description: What test, cookbook, or release gate should prove this works?
+    validations:
+      required: false

.github/PULL_REQUEST_TEMPLATE.md

@@ -1,28 +1,19 @@
-## What does this PR do?
+## Summary
 
-<!-- One paragraph. What problem does it solve, or what capability does it add? -->
+Describe the user-facing change and why it is needed.
 
-## Why?
+## Verification
 
-<!-- What was broken or missing? Link to any related issue: "Fixes #123" or "Closes #456" -->
-
-## How was it tested?
-
-<!-- Which tests cover this change? Did you add new ones? Did you run the integration tests? -->
-
-- [ ] Unit tests added / updated (`pytest` or `vitest`)
-- [ ] Integration tests pass (or N/A — no gateway behavior changed)
-- [ ] `ruff check` / `mypy` / `npm run typecheck` / `npm run lint` all pass
-- [ ] Public types, env vars, or SDK behavior changes are documented in the relevant README
+- [ ] `uv run ruff check .`
+- [ ] Focused pytest target:
+- [ ] Full `uv run pytest -q` when touching shared runtime, CLI, SDK, or release gates
+- [ ] TypeScript build/test when touching `typescript/`
+- [ ] `agent-learn release-proof` when touching release-candidate behavior
 
 ## Checklist
 
-- [ ] Branch is off `main`
-- [ ] Commit messages follow [Conventional Commits](https://www.conventionalcommits.org/) (`feat:`, `fix:`, `docs:`, `chore:` …)
-- [ ] No TODOs or commented-out code left in
-- [ ] No real API keys or secrets in the diff
-- [ ] If prose was added or changed: checked against [`docs/VOCABULARY.md`](../docs/VOCABULARY.md)
-
-## Notes for reviewers
-
-<!-- Anything that needs extra attention, or context that isn't obvious from the diff. -->
+- [ ] Public API names stay under `agent_learning.*`, `agent-learn`, or `@future-agi/agent-learning-kit`
+- [ ] New user-facing workflow has docs or an example
+- [ ] New release claim has executable gate coverage
+- [ ] No hosted-service dependency was added to release-gated examples
+- [ ] Security-sensitive output is redacted or justified