Developer shares full trace of local AI agent code review with an 'armed fence' approval workflow
HankYeomans · x · 2026-10-14
HankYeomans shares a sequence diagram of one review-and-bug-fix cycle on his local AI instance. The workflow centers on an 'armed fence': after task approval, jevscopearm dry-runs, records a pending approval to .jev-dev/pending-arm-approval., then consumes an operator /jev approve command and emits an armed receipt (JSON). The agent then edits within fenced paths, jevcheckpoint fires once per round, and a Reviewer model (gpt-6-astra/high) recaptures the snapshot, verifies the fence, reruns checks, and returns a bound review packet. Clean runs are marked reviewed; findings within budget trigger a correction prompt at the stop boundary followed by a fix and another checkpoint. The author admits the loop is still clunky with him in it for demos.
More from coding & agent
- SlashVibe, a physical controller for AI agents, hits Kickstarter October 27 — sujingshen · 2026-10-14
- Rosebud AI lets coding agents publish your browser game with free hosting and a play link — lishali88 · 2026-10-14
- Claude Code 2.1.300 Is About to Drop, No Changelog Yet — ClaudeCodeLog · 2026-10-14
- Epic releases AI-assisted launcher preview: official Linux support and 1.8x faster startup than Steam — AIFlow_ML · 2026-10-14
- a16z: Long-Running Agents Take Over Relay-Race Workflows, Pricing Shifts to Outcomes — a16z · 2026-10-14
- Claude builds its own test tool and fixes Meta VR depth occlusion flicker in minutes — nptacek · 2026-10-14