How do you test shared agent skills across repos? A dev asks for frameworks
augburto · reddit · 2026-09-15
A developer describes a common agent-engineering problem: their company keeps a central repository of generic, cross-repo agent skills, some of which are operational. When updating these skills, they want a way to verify quality hasn't regressed.
Their instinct is to build evals against a gold standard, but that feels heavy-handed and they're asking whether the community has simpler, battle-tested approaches for testing shared skills.
More from coding & agent
- Replit launches Routines: scheduled production monitoring with agent investigation — amasad · 2026-09-15
- Hackathon Project Skillio Turns Vision Pro Into an AR Screw-Driving Coach Agent — seanmcdonaldxyz · 2026-09-15
- Let the Model Propose, the Flow Validate, a Human Approve: Drafting Agent Checkpoints — WirelessLife · 2026-09-15
- Devin-powered daily briefs: how one dev uses AI agents to automate information discovery — bendee983 · 2026-09-15
- Claude writes 80% of code at Anthropic, CI jobs up 25x in six months — addyosmani · 2026-09-15
- Voice agent production postmortem: 7 behavioral bugs found only in real call logs — authentic_developer · 2026-09-15