Show HN: AgentToolBench-Code – security benchmark for AI coding agents
The AgentToolBench-Code has expanded its security benchmark for AI coding agents from 10 to 16 scenarios. The results indicate that Sonnet 4.6 outperformed Haiku 4.5 in identifying silent security failures. This benchmark aims to provide insights into the capabilities of AI coding agents in real-world scenarios.
- ▪The benchmark tests eight scoring axes with two scenarios each, anchored to real-world coding-agent attack classes.
- ▪Sonnet 4.6 scored +9 out of 16, while Haiku 4.5 scored +3 out of 16 in the expanded corpus.
- ▪The benchmark revealed that Sonnet catches more vulnerabilities, including PyPI typosquats and internal IPs, which Haiku missed.
Hacker News (AI / LLM) files mainly under ai. We currently carry 3,338 of its stories.
Story provenance
Source · retrieval · rights · ranking — open for full record
inspect →
Story provenance
Attribution is not the same as permission. This drawer separates discovery metadata, excerpts, WeSearch-generated summaries, reuse status, and whether the publisher receives the visit. Nothing here claims a legal grant the publisher has not made.
Record
| Original publisher | Gist |
| Canonical URL | https://gist.github.com/allenwu-blip/fa2bd0218b93a1d7aef765817e3c6608 |
| Publication time | Tue, 26 May 2026 03:45:20 +0000 |
| Retrieval time | 2026-05-26T03:47:42.974Z |
| Last seen | 2026-05-26T03:47:42.974Z |
| Headline source | Publisher (no WeSearch rewrite) |
| Excerpt source | publisher body |
| Excerpt method | First ~120 words (~800 chars) of extracted publisher body, fair-use limited. |
| Summary | WeSearch · cerebras-chat (WeSearch summarizer) |
| Summary source text | contentText |
| Citation coverage | Summary is a WeSearch-generated derivative; primary citation is the original publisher URL. |
| Cluster | WCFjipAG9JTd |
| Cluster logic | Grouped by semantic title/content similarity across sources within a rolling window. Same-publisher template collisions are excluded from coverage comparison. |
| Ranking reason | Story pages are not engagement-ranked. Hub feeds use recency, with optional source-diversified chronological ordering (cap consecutive stories per source). No personalized ranking. |
| Publisher visit | Yes — open original |
| Substitutes article? | No — link-out required for full text |
Rights status (four layers)
WeSearch handling by dimension
| Indexing | May the item be indexed (stored, ranked, made findable)? | Allowed |
| Snippet | May a short excerpt of the publisher's text be shown? | Allowed |
| AI summary | May WeSearch generate its own short summary of the article? | Limited |
| Retrieval / RAG | May the content be exposed for third-party retrieval-augmented generation? | Not asserted |
| Model training | May the content be used to train AI models? | Not asserted |
| Commercial reuse | May the content be reused commercially? | Not permitted |
Basis: Derived from the published RSS/Atom feed. Contact: [email protected]. Reviewed: 2026-07-24.
Opening excerpt (first ~120 words) tap to expand
I doubled my AI-agent security benchmark from 10 scenarios to 16. The "Sonnet vs Haiku tie" disappeared. Draft launch post for AgentToolBench-Code v0.0.1 — not yet published. All numbers verified against examples/claude-code-sonnet-16.jsonl and examples/claude-code-haiku-16.jsonl in the repo. Re-runnable from a clean checkout for ~$4 of Anthropic API. A week ago I shipped v0.0.1 of AgentToolBench-Code, an open-source benchmark for silent security failures in AI coding agents. The first empirical finding — that Claude Code Sonnet 4.6 and Haiku 4.5 scored identically (+5/+10) on a 10-scenario corpus — was striking enough that I wrote it up.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Gist.
Discussion
0 commentsMore from Gist
