At-Risk Detection
Proactively scan GitHub issues and pull requests to identify at-risk items before they become problems. This skill detects stale items, blocked work, deadline risks, scope creep, and unassigned high-priority issues, then provides actionable suggestions for each.
Installation
Add the marketplace and install this skill:
/plugin marketplace add britt/claude-code-skills
/plugin install at-risk-detection@britt
Or install all skills at once:
/plugin install claude-code-skills@britt
Compatibility
| Platform | Supported |
|---|---|
| Claude Code | ✓ |
| Claude.ai | ✓ |
How to use it
Use this skill when:
- Starting a new sprint or running a weekly health check
- During standup to identify blockers
- Checking project health or status before milestone deadlines
- Triaging backlog items or doing pre-sprint cleanup
Features
Multi-signal risk detection Detects five risk types: stale items, blocked work, deadline risks, scope creep, and assignment issues using configurable heuristics and thresholds.
Health score and summary reports Produces a project health summary with total open items, at-risk counts, health score percentage, and risk breakdown by severity.
Actionable suggestions For each risk signal detected, provides specific recommendations such as commenting with status updates, resolving blocking dependencies, or redistributing overloaded assignees.
Quick and comprehensive modes Quick mode scans the 30 most recent open issues for daily checks; comprehensive mode paginates through all open issues for thorough milestone reviews. Because quick mode sorts by recency, it can miss the stalest items — the report calls this out, and comprehensive mode should be used to find them.
Common-mistake guardrails
Avoids known false positives: draft PRs aren’t flagged as stale-awaiting-review, and issues intentionally parked with an on-hold label are reported separately from genuinely forgotten stale work.