<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Mind the GAP: Text Safety Does Not Transfer to Tool-Call Safety in LLM Agents · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2602.16943/</link><description>Collected HN comments and submissions linking this paper.</description><language>en</language><lastBuildDate>Mon, 02 Nov 2026 11:27:26 GMT</lastBuildDate><atom:link href="https://hackerpaper.kaipfister.de/p/2602.16943/feed.xml" rel="self" type="application/rss+xml" /><item><title>Mind the GAP: Text Safety Does Not Transfer to Tool-Call Safety in LLM Agents — HN submission</title><link>https://news.ycombinator.com/item?id=47159542</link><guid isPermaLink="false">urn:hackerpapers:submission:2602.16943:47159542</guid><pubDate>Wed, 25 Feb 2026 23:27:52 GMT</pubDate><description>&lt;p&gt;Show HN: Edictum – Runtime governance for LLM agent tool calls&lt;/p&gt;&lt;p&gt;We tested 6 frontier models across 17,420 tool-call interactions and found that models consistently refuse harmful requests in text while executing them through tool calls. We call this divergence the GAP metric. The text says no. The tool call says yes. Edictum is a runtime governance library that enforces safety contracts at the tool-call boundary — the point where you have the tool name, the arguments, and the ability to block before execution. YAML contracts with preconditions, postconditions, PII redaction. Deterministic allow/deny/redact, no LLM-in-the-loop. Zero runtime dependencies…&lt;/p&gt;&lt;p&gt;2 points; 0 discussion replies on HN. &lt;a href="https://hackerpaper.kaipfister.de/p/2602.16943/"&gt;Paper and source history&lt;/a&gt;. A submission is not an endorsement.&lt;/p&gt;</description><category>cs.AI</category></item></channel></rss>
