Steering coding agents in large repos: missing guardrails produce more slop
yoyoyoav · reddit · 2026-10-14
Reddit user yoyoyoav shares experiments on making coding agents automatically pick the right context or skill instead of users choosing tools manually. In large repos where everyone does things differently, agents lack guardrails and produce far more slop than in personal projects. The post asks the community for tools and tips on steering agents toward efficiency and correctness.
More from coding & agent
- Claude builds its own test tool, fixes DreamPark MR sky-flicker bug in minutes — nptacek · 2026-10-14
- Argon arrives in Google Antigravity but users say it's useless even on the 5x plan — waqasalibrand · 2026-10-14
- CMU paper: a 4B model rewrites an agent's harness and beats the 35B model it learned from — irinarish · 2026-10-14
- Harness ≠ infrastructure: sandboxing and persistence belong to a different layer — blaizedsouza · 2026-10-14
- Anthropic publishes security guide for locking down Claude Managed Agents sandboxes — brada · 2026-10-14
- Dev used Gemini as software architect to build an Apple Watch caffeine tracker — pythononrailz · 2026-10-14