Convex Insights

Convex-insights is for the moment you're staring at a live Convex app and want to know what's actually happening without clicking through the dashboard. You ask in plain language ("what's erroring since the last deploy?", "why did this call fail?") and it runs narrow read-only queries against the official Convex MCP, filtering logs client-side because the underlying tools have no server-side filters. Answers come back as a one-line finding, counts, and one representative stack, plus a dashboard link to verify. The honest caveat: it's a disciplined reader, not a fixer. Perf and cost root-causing gets handed to convex-advisor, and it only sees the last bounded window of logs, so very high-volume apps need a tighter question.

skills.sh installs: 41K
GitHub stars: 69
last updated: Oct 26, 2026

View on GitHub

Install

$ npx -y skills add convex/agent-skills --skill convex-insights --agent claude-code

Installs into .claude/skills of the current project.

Files

Query logs + health in natural language

The deployment already records what happened; the agent just has to ask well. This capability is a disciplined wrapper over the official Convex MCP's read tools (logs, insights, functionSpec, status) that turns operational questions into narrow, evidence-returning queries and hands back answers a human can one-click verify in the dashboard. The discipline is copied from the observability MCP surface that works best in the wild: discover fields before querying, three views not fifteen tools, token-frugal output, and a dashboard deep link on every answer.