Azure / Azure/azure-sdk-tools

Detect teams struggling with default process and recommend library-specific skills

Open
#14,733 0 comments 0 reactions 0 assignees View on GitHub
AzSDK Tools Agent
Dominant language
C#
Stars
135
Forks
260
Avg merge
1d 22h
Merged PRs (30d)
115

Description

## Description

Add logic to the Azure SDK Tools Agent that detects when a service team is having repeated issues or friction with the default process (e.g., failing steps, unusual workarounds, repeated troubleshooting). When the agent identifies that a team has unique steps or deviations from the standard workflow, it should proactively recommend that the team create a Copilot skill specific to their library.

This enables teams with non-standard inner-loop requirements to codify their unique process into a reusable skill, improving their experience and reducing friction over time.

### Expected Behavior

1. **Detection** — Agent monitors for signals that a team's workflow diverges from the default process (repeated errors, manual workarounds, skipped steps, unique build/test configurations).
2. **Recommendation** — When divergence is detected, the agent suggests creating a library-specific skill and provides guidance on how to get started (linking to skill-authoring resources).
3. **Follow-through** — Optionally offer to scaffold the initial skill structure for the team.

Contributor guide

Open the contributing guide

Research direction

Start by locating the Azure SDK Tools Agent workflow and the skill-authoring resources referenced by the issue. Trace where workflow outcomes, errors, workarounds, and configuration deviations are available, then determine how recommendations and optional scaffolding fit that flow. Done means repeated divergence can lead to a library-specific skill recommendation with authoring guidance.

Written by the indexing model from the issue text.

Assessment

Tech stack
azure, csharp
Domain
ai, tooling
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
42/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.