Name: microsoft-foundry
Rating: 63.2 (1 reviews)
Author: microsoft

microsoft-foundry

Deploy, evaluate, fine-tune, and manage Foundry agents end-to-end with azd: hosted agent scaffold/run/deploy, prompt agent create, batch eval, continuous eval, prompt optimizer, Agent Optimizer scaffold, agent.yaml, dataset curation from traces, model fine-tuning (SFT/DPO/RFT). USE FOR: azd ai agent, azd provision/deploy, deploy agent, hosted agent, create agent, add tool to agent, invoke agent, evaluate agent, continuous eval, continuous monitoring, optimize prompt, improve prompt, optimize agent instructions, agent optimizer, deploy model, Foundry project, RBAC, role assignment, permissions, quota, capacity, region, troubleshoot agent, deployment failure, AI Services, create Foundry resource, provision, knowledge index, customize deployment, onboard, availability, fine-tune, SFT, DPO, RFT, training-data, grader, distillation, fine-tuned model, large file upload. DO NOT USE FOR: Azure Functions, App Service, general Azure deploy (use azure-deploy), general Azure prep (use azure-prepare).

Quality

73%

Does it follow best practices?

Impact

—

No eval scenarios have been run

Securityby

Failed to scan

The risk profile of this skill

Fix and improve this skill with Tessl

tessl review fix ./plugin/skills/microsoft-foundry/SKILL.md

Quality

Content

47%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

This skill functions well as a comprehensive routing hub with excellent workflow clarity and decision tables for matching user intent to sub-skills. However, it suffers significantly from verbosity—the 7-step context resolution procedure, metadata overlay rules, and legacy migration logic bloat the main file far beyond what's needed for routing. Moving these procedural details to a referenced sub-document would dramatically improve conciseness and progressive disclosure.

Suggestions

Extract Steps 1-7 ('Common Project Context Resolution') and the metadata write rules into a separate reference document (e.g., 'references/context-resolution.md') and replace with a 2-3 line summary and link in the main SKILL.md.

Remove or relocate the '.foundry Workspace Standard' section and 'Setup References' section to referenced documents—they are implementation details that don't belong in the top-level routing skill.

Add at least one concrete executable example (e.g., a sample azd command or MCP tool invocation) to the main skill body to improve actionability beyond pure routing tables.

Trim the legacy metadata migration logic (testSuites/testCases normalization, priority-to-tags mapping) into a dedicated migration reference document rather than embedding it in the main routing skill.

Dimension	Reasoning	Score
Conciseness	The skill is extremely verbose at ~300+ lines. Much of the content—especially the multi-step context resolution (Steps 1-7), metadata overlay rules, legacy migration logic, and environment resolution cascades—is exhaustive procedural detail that could be in a referenced sub-document. The main SKILL.md reads like an internal specification rather than a lean routing document.	1 / 3
Actionability	The skill provides concrete routing tables, specific azd variable names, and clear MCP tool references, which is useful. However, it contains no executable code examples or copy-paste-ready commands in the main body—most actionable content is deferred to sub-skills. The context resolution steps are detailed but procedural/descriptive rather than executable.	2 / 3
Workflow Clarity	The multi-step context resolution workflow (Steps 1-7) is clearly sequenced with explicit decision points, fallback logic, and validation checkpoints (e.g., 'stop and ask which source is authoritative' when azd and metadata conflict). The agent development lifecycle table provides clear workflow chains with ordering. The infrastructure lifecycle includes disambiguation steps.	3 / 3
Progressive Disclosure	The skill correctly uses a sub-skill table with clear routing and references to 17+ sub-documents, which is good progressive disclosure architecture. However, the main SKILL.md itself is a wall of text—Steps 1-7 of 'Common Project Context Resolution' and the metadata overlay rules are inline content that should be in a referenced document (e.g., 'references/context-resolution.md'), keeping the main skill as a lean routing hub.	2 / 3
	Total	8 / 12 Passed

Description

100%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

This is a strong skill description that excels across all dimensions. It provides comprehensive specific actions, extensive natural trigger terms via the USE FOR section, clear what/when guidance, and explicit disambiguation from related skills via the DO NOT USE FOR clause. The only minor weakness is that the density of terms makes it somewhat hard to parse for a human reader, but for Claude's skill selection purposes it is highly effective.

Dimension	Reasoning	Score
Specificity	Lists multiple specific concrete actions: deploy, evaluate, fine-tune, manage agents, scaffold/run/deploy hosted agents, prompt agent create, batch eval, continuous eval, prompt optimizer, Agent Optimizer scaffold, dataset curation from traces, model fine-tuning with specific methods (SFT/DPO/RFT).	3 / 3
Completeness	Clearly answers both 'what' (deploy, evaluate, fine-tune, manage Foundry agents with azd) and 'when' (explicit USE FOR clause with extensive trigger terms, plus a DO NOT USE FOR clause that further clarifies boundaries).	3 / 3
Trigger Term Quality	Excellent coverage of natural terms users would say, including 'deploy agent', 'create agent', 'evaluate agent', 'optimize prompt', 'fine-tune', 'troubleshoot agent', 'deployment failure', 'add tool to agent', and many more variations. The USE FOR section provides comprehensive keyword coverage.	3 / 3
Distinctiveness Conflict Risk	Highly distinctive with clear niche around Foundry agents and azd tooling. The DO NOT USE FOR clause explicitly disambiguates from azure-deploy and azure-prepare skills, reducing conflict risk significantly.	3 / 3
	Total	12 / 12 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 11 / 11 Passed

Validation for skill structure

No warnings or errors.

Repository: microsoft/github-copilot-for-azure
Commit: d4563f6

Reviewed: 3 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.