Torturing agents with 10x test harnesses: a workflow for constraining coding AI
cjimti · x · 2026-09-24
Developer @cjimti describes his approach to agentic coding: he adds 10x more test harnesses than any human team would tolerate, including mutation testing, deliberately to slow agents down. In his regime, any test failure is treated as a process failure requiring full review, and agents are penalized for using the test suite as their debug tool — a test-density-first philosophy for constraining autonomous coders.
Related event: Developer Slows AI Agents with 10x More Test Harnesses(2 posts)→
More from coding & agent
- Pigment Agents now reads and posts in Slack, inheriting user permissions — mattturck · 2026-11-12
- Map of the agent economy: 50+ startups across payments, spend control, identity, memory — venturetwins · 2026-11-12
- OpenCode launches Browse, an on-device browser agent with local logins — Saul_Loveman · 2026-11-12
- VentureTwins maps the B2A economy: dozens of agent infra and app companies in first series installment — venturetwins · 2026-11-12
- Google releases 76-page 'Agents Companion' whitepaper on agent engineering — mdancho84 · 2026-11-12
- Spinning up containers for 20-line agent tools is backwards: Safescript compiles provable, 50ms-deploy apps — uriwa · 2026-11-12