safeagents plugs decision models into the loop, auditing only suspicious agent calls
yb2698 · x · 2026-10-29
safeagents defaults to rule-based checks, but a simple argument lets you plug any decision-model API into the loop for stronger inspection.
- Why decision models fit: feedback is fast — the model returns only a short JSON verdict (allow/block plus one sentence), and it's only asked about calls the rules already flagged, so clean calls cost nothing.
- Compact, unbribable prompt: it only sees the user request, the tool call, and atif-scan findings — never the agent's reasoning or tool outputs — so the agent can't argue its way past it, and a small local model suffices.
Related event: safeagents adds pluggable decision models and log-only mode(2 posts)→
More from coding & agent
- Synap claims #1 on LongMemEval with 95.6% answer accuracy, optimizing a different metric — arpit_bhayani · 2026-10-29
- Min-cut on a 4.49M-transfer graph to freeze mule accounts (and a max-flow bug) — JeremyCMorgan · 2026-10-29
- Dev admits 'dangerously bypassing permissions' with coding agents from day one — brandon_galang · 2026-10-29
- Ultracode in Claude Code isn't a thinking level — it splits tasks across subagents — jacob_posel · 2026-10-29
- Hack lets any Grok bot switch to Opus by routing requests through a Cursor project — bfrench · 2026-10-29
- Director's Cut: open-source ComfyUI filmmaking app with character presets — oLeonardoFrito · 2026-10-29