MCP Prompts

adtk registers 4 MCP prompts that guide AI agents through multi-step analysis workflows. Prompts are templates that instruct the agent which tools to call and how to present the results.

sprint_summary

Generate a sprint/iteration summary for an Azure DevOps project.

Arguments

Argument	Required	Description
`project`	Yes	Azure DevOps project name
`team`	No	Team name (scopes to a specific team)
`iteration`	No	Iteration path or name (defaults to current iteration)

What it does

Fetches the current (or specified) iteration via manage_iterations
Gets all work items in the iteration via manage_work_items → iteration_items
Gets the board workflow via manage_boards → get_columns
Categorizes work items by state and type

Output format

Sprint name, dates, and timeline progress
Completion percentage (done / total)
Work items grouped by state
Breakdown by work item type
Items at risk (active but not updated recently)
Blocked items or impediments
Team velocity summary

Example usage

Prompt: sprint_summary
Arguments: { "project": "MyProject", "team": "Backend Team" }

pr_review_digest

Generate a PR review digest for an Azure DevOps repository.

Arguments

Argument	Required	Description
`project`	Yes	Azure DevOps project name
`repo`	Yes	Repository name

What it does

Lists all active PRs via manage_pull_requests → list
Checks reviewer status for each PR via manage_pull_requests → list_reviewers
Analyzes review status, age, and potential issues

Output format

Total open PRs and summary statistics
PRs needing review (no reviewers or pending votes)
PRs with requested changes (rejected votes)
PRs ready to merge (all approved)
Stale PRs (open > 7 days without activity)
PRs by author showing review workload distribution

Example usage

Prompt: pr_review_digest
Arguments: { "project": "MyProject", "repo": "backend-api" }

pipeline_health

Analyze CI/CD pipeline health for an Azure DevOps project.

Arguments

Argument	Required	Description
`project`	Yes	Azure DevOps project name
`pipeline_id`	No	Specific pipeline ID to analyze (omit for all pipelines)

What it does

Lists pipelines (or gets a specific one) via manage_pipelines
Gets recent runs via manage_pipelines → list_runs (top 20)
For failures, inspects run details via manage_pipelines → get_run
Optionally inspects build logs via manage_pipelines → get_logs

Output format

Overall health score (pass rate)
Success/failure rate breakdown
Average build duration and trends
Most common failure reasons
Pipelines with consecutive failures
Longest running pipelines
Recommendations for improving reliability

Example usage

Prompt: pipeline_health
Arguments: { "project": "MyProject", "pipeline_id": "42" }

release_readiness

Assess release readiness for an Azure DevOps project.

Arguments

Argument	Required	Description
`project`	Yes	Azure DevOps project name
`iteration`	No	Iteration path or name (defaults to current iteration)

What it does

Fetches the current (or specified) iteration via manage_iterations
Gets all work items in the iteration via manage_work_items → iteration_items
Queries open bugs via manage_search → wiql
Lists unmerged PRs via manage_pull_requests → list
Checks recent build status via manage_pipelines → list_runs

Output format

Overall readiness score (Ready / At Risk / Not Ready)
Work item completion: completed vs remaining
Open bugs: count and severity breakdown
Unmerged PRs: list with status and blockers
Build status: latest pipeline results
Test results from recent builds
Blockers and risks
Go/No-Go recommendation with justification

Example usage

Prompt: release_readiness
Arguments: { "project": "MyProject", "iteration": "Sprint 24" }