vincentvandeth

5 ms·

1.
Show HN: Glass box governance for multi-agent AI coding workflows(github.com)
1 points by vincentvandeth 7mo ago|0 comments
2.
by vincentvandeth 7mo ago
Update: I ended up building this into a full closed-loop pipeline. A PreToolUse hook detects context pressure at 65%, the agent writes a structured handover (task state, files, progress), tmux clears the session, and a rotator script inject
3.
Show HN: Automatic context rotation for Claude Code (no manual steps)
2 points by vincentvandeth 7mo ago|0 comments
4.
by vincentvandeth 7mo ago
Nice — just checked it out. The interceptor approach makes sense for teams that need policy enforcement across multiple agents. Interesting difference in philosophy though: Limits enforces rules defined upfront, while what I built learns ru
5.
by vincentvandeth 7mo ago
Fair point on the time cost — but I'd frame it differently. The 6 months wasn't spent building a governance layer instead of building the agent. The governance layer grew out of the actual project work. Every receipt, every qualit
6.
I built a governance layer for multi-agent AI coding – lessons after 6 months
2 points by vincentvandeth 7mo ago|0 comments
7.
by vincentvandeth 7mo ago
Context bloat is a real problem — and yes, order matters more than most people realize. Descriptions near the top of the tool list get preferentially selected, especially in long contexts where attention degrades. Two things I do to fight t
8.
by vincentvandeth 7mo ago
About 6 months of iterating, but in bursts — I built it while using it on a production project, so the governance layer grew alongside real failure modes rather than being designed upfront. The maintenance question is the right one. The rul
9.
by vincentvandeth 7mo ago
Hard-coded checks before every action, plus a governance layer that separates "what the agent wants to do" from "what it's allowed to do." The deeper issue: if your agent decides whether to issue a refund, you'
10.
by vincentvandeth 7mo ago
Yes, tmux. The setup is a 2x2 grid: T0 (orchestrator) | T1 (Track A) T2 (Track B) | T3 (Track C) When a worker finishes, it writes a structured report to a shared unified_reports/ directory. A file watcher (receipt processor)
11.
by vincentvandeth 7mo ago
You're right that no model handles everything in one context window — that's exactly why I built context rotation. Each task runs in a single agent context (one responsibility, clear scope), and when the window fills up, the syste
12.
by vincentvandeth 7mo ago
This approach sounds clean in theory, but in production you're building a black box. When your planning agent hands off to an implementation agent and that hands off to a review agent — where did the bug originate? Which agent's c
13.
Ask HN: Anyone running AI agents in production without engineering background?
2 points by vincentvandeth 7mo ago|0 comments
14.
by vincentvandeth 7mo ago
Great list. I've been running a multi-agent orchestration system (11 specialized AI agents) in production for 6 months and your #2 and #5 resonate hard. What I'd add: 6. Confidence without evidence. Agents will report "task c
15.
by vincentvandeth 7mo ago
Great questions. I've been running ~2,400 multi-agent dispatches across 4 terminals (different AI models) for about 6 months, so I'll share what I've hit in practice rather than theory. On RFC 3161 vs. multi-witness anchoring
16.
by vincentvandeth 7mo ago
Interesting approach. Runtime enforcement is the part most people skip — they focus on logging what happened but don't prevent bad actions in the first place. The policy engine + kill switch combination makes sense for that. I've