Agent-designed specs tested against every historical CVE, then adversarially attacked
jd_pressman · x · 2026-10-24
jdpressman outlines a workflow for AI-generated software specs: let the agent pick the spec, walk it through every historical CVE to check whether the spec would in principle have stopped each vulnerability, then run an adversarial pass in breaker mode to find weaknesses. Replying to a note that even lazy specs deliver outsized value, he calls it a 90/10 problem — any adversarial pass captures most of the benefit, with deeper review done by asking the agent for a guided codebase tour while taking direct control of its attention.
Related event: Testing Agent-Written Specs by Replaying Historical CVEs Against Them(4 posts)→
More from coding & agent
- Cut LLM costs: route requests, disable thinking, cache answers, escalate only on low confidence — colinmcnamara · 2026-10-24
- NVIDIA's OpenShell: Apache-2.0 runtime sandbox for autonomous agent fleets, 18.6k stars — petrusenko_max · 2026-10-24
- Claude Code Routines runs overnight, screens 380 clinics into 30 outreach drafts — Arindam_1729 · 2026-10-24
- Factory and Anthropic share customer story on long-horizon coding agents and 'dark factories' — matanSF · 2026-10-24
- Dev argues Rust rewrite of Apache2 with Verus spec beats running C code in expectation — jd_pressman · 2026-10-24
- Carmack amplifies Goose: a memory-safe systems language that beats C++ and Rust without a heap — ID_AA_Carmack · 2026-10-24