Skip to content

Optimize FAOS preparation and clarify evaluator documentation - #2482

Merged
Luffy Chen (XOEEst) merged 32 commits into
microsoft:mainfrom
XOEEst:main
Jun 2, 2026
Merged

Optimize FAOS preparation and clarify evaluator documentation#2482
Luffy Chen (XOEEst) merged 32 commits into
microsoft:mainfrom
XOEEst:main

Conversation

@XOEEst

Copy link
Copy Markdown
Contributor

This pull request updates the documentation and test snapshots for the Microsoft Foundry skill, focusing on clarifying terminology, updating SDK installation instructions, and enforcing naming conventions for evaluation suites. The most significant changes are grouped below:

SDK Installation and Usage Updates:

  • Updated all documentation to recommend installing the azure-ai-agentserver-optimization package directly via PyPI, removing instructions for using a local preview wheel. This simplifies setup and ensures consistency as the package is now publicly available. [1] [2] [3]

Evaluator and Evaluation Suite Terminology:

  • Replaced references to "adaptive evaluator" with "rubric-based evaluator" throughout the documentation for clarity and consistency. [1] [2] [3] [4] [5] [6]
  • Clarified that evaluation suite names (suiteName) must start with a letter, and provided guidance for prefixing numeric-derived names with an alphabetic label (e.g., suite-). [1] [2] [3]

Skill Description and Trigger Keywords:

  • Updated the skill description in SKILL.md and corresponding test snapshots to reflect new capabilities (e.g., "monitoring", "agent optimizer", "invocations WebSocket") and removed deprecated or redundant terms. [1] [2]
  • Adjusted the extracted trigger keywords in test snapshots to match the revised description and terminology. [1] [2] [3] [4] [5]## Description

Checklist

  • Tests pass locally (cd tests && npm test)
  • If modifying skill descriptions: verified routing correctness with integration tests (npm run test:skills:integration -- <skill>)
  • If modifying skill USE FOR / DO NOT USE FOR / PREFER OVER clauses: confirmed no routing regressions for competing skills

Related Issues

Luffy Chen (XOEEst) and others added 30 commits May 5, 2026 22:38
Address PR microsoft#2174 review comments by expanding FAOS as Foundry Agent Optimization Service on first use and making the Step 8 Python config snippet copy-paste safe.

The minimum contract example now imports os and preserves the app's existing model-selection fallback instead of hard-coding MODEL_DEPLOYMENT_NAME unless that is already what the app uses.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
# Conflicts:
#	plugin/skills/microsoft-foundry/SKILL.md
When creating custom evaluator prompts, treat the runtime-enforced JSON schema as authoritative: result plus reason. Preserve user-provided rubric text, but remove or normalize conflicting output schemas such as score/reasoning or duplicate OUTPUT FORMAT blocks before calling evaluator_catalog_create.

Add observe skill test coverage for the promptText guardrail so future edits keep the result/reason contract visible.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Remove the 0-to-1 scoringType/minScore/maxScore details from the sample custom evaluator prompt. The important guardrail is the output contract: preserve the rubric, but avoid conflicting output schemas because the runtime enforces result/reason.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Restore plugin/skills/microsoft-foundry/SKILL.md from upstream/main so this PR no longer carries a line-ending-only change for the top-level skill file.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Updates the Microsoft Foundry skill documentation and its trigger keyword snapshots to reflect current SDK installation guidance and clearer evaluation terminology, while tightening evaluation-suite naming guidance for consistency with expected service constraints.

Changes:

  • Updated the microsoft-foundry skill description (and corresponding trigger keyword snapshots) to reflect revised capabilities/keywords (notably removing “container start” and adding “agent optimizer”).
  • Replaced “adaptive evaluator” wording with “rubric-based evaluator” across the observability/evaluation docs.
  • Simplified Agent Optimizer setup docs to use the published PyPI package azure-ai-agentserver-optimization and added guidance that suiteName must start with a letter.

Reviewed changes

Copilot reviewed 23 out of 23 changed files in this pull request and generated no comments.

Show a summary per file
File Description
tests/microsoft-foundry/resource/create/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/models/deploy/deploy-model/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/models/deploy/deploy-model-optimal-region/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/models/deploy/customize-deployment/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/models/deploy/capacity/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/foundry-agent/troubleshoot/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/foundry-agent/trace/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/foundry-agent/observe/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/foundry-agent/invoke/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/foundry-agent/eval-datasets/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/foundry-agent/deploy/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/foundry-agent/create/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/finetuning/snapshots/triggers.test.ts.snap Updates trigger snapshot to match revised skill description/keywords.
tests/microsoft-foundry/snapshots/triggers.test.ts.snap Updates top-level trigger snapshot to match revised skill description/keywords.
plugin/skills/microsoft-foundry/SKILL.md Updates the skill frontmatter description to align with updated terminology/capabilities and expected routing keywords.
plugin/skills/microsoft-foundry/references/agent-metadata-contract.md Adds explicit suite naming guidance (suiteName must start with a letter).
plugin/skills/microsoft-foundry/foundry-agent/observe/references/evaluation-suite-generation.md Renames evaluator terminology and documents suite naming constraints for generated suites.
plugin/skills/microsoft-foundry/foundry-agent/observe/references/evaluate-step.md Updates evaluator terminology to “rubric-based”.
plugin/skills/microsoft-foundry/foundry-agent/observe/references/deploy-and-setup.md Documents suite naming constraints during auto-setup.
plugin/skills/microsoft-foundry/foundry-agent/observe/observe.md Updates evaluator terminology across the observability workflow description/guardrails.
plugin/skills/microsoft-foundry/foundry-agent/agent-optimizer/references/scaffold.md Switches install guidance from preview wheel to PyPI package.
plugin/skills/microsoft-foundry/foundry-agent/agent-optimizer/references/python-patterns.md Switches install guidance from preview wheel to PyPI package and updates verification checklist.
plugin/skills/microsoft-foundry/foundry-agent/agent-optimizer/references/eval-yaml.md Clarifies/normalizes eval.yaml example structure and adds guidance to avoid including the baseline model as a target candidate.

@XOEEst
Luffy Chen (XOEEst) merged commit 3890cbf into microsoft:main Jun 2, 2026
12 checks passed
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.

3 participants