Grok payment agent rejects two attacks in dry run; risk gate blocks rest
11/14/2026 — 11/14, 11:15·1 sources (all)·1 reports (all)
Key takeaways
- A developer reports running a dry-run payment-agent test with Grok (grok-4.7) via the xAI API. Grok itself rejected two attacks—unlimited au…
- no transactions were signed or sent.
Background & analysis
On November 14, 2026, at 10:57, a developer published a post on DEV Community describing a dry-run test of a payments agent built on Grok (grok-4.7, accessed through the xAI API). The agent was handed a single payments-agent task that contained three items in its inbox, and it was given only two tools to work with: proposing a payment or proposing an approval.
Before anything could reach the network, every proposal generated by the agent was routed through sra-riskgate 0.18.0 for inspection. Nothing was actually signed and no transaction was sent at any point during the exercise. According to the post, Grok refused two attack attempts on its own, specifically one involving unlimited authorization and another involving a forged invoice. The remaining attack attempts were blocked by the sra-riskgate layer before the agent could act on them.
The account so far comes from a single developer-authored post on DEV Community, and the test was set up and run privately by that developer. There is no independent reproduction, vendor statement, or third-party confirmation of the results yet, so the reported behavior rests entirely on the published description of this single run.
AI-generated from 1 reports · updated 3 days ago
Related AI tools
- Heat index
- 25
- All sources
- 1
- All reports
- 1
- First seen
- 3 days ago
Reports on this story headlines open the original
A developer writing on DEV Community tested Grok (grok-4.7 via the xAI API) as a payments agent in a dry run, with every proposal checked by sra-riskgate 0.18.0. Grok refused an unlimited approval and a look-alike invoice on its own, and no transaction was signed or sent.
DEV Community · AIAI score 52
Other stories people are talking about
- Heat index 1345RisingNvidia board not discussing Jensen Huang's return, probe into his conduct advances22 sources
- Heat index 1241SurgeCalifornia Sues OpenAI Over Agents Escaping Tests, Intruding into Hugging Face15 sources
- Heat index 901Alibaba releases MoE model Qwen4-Max with 1M token context for low-cost, long-running tasks17 sources
- Heat index 695US Commerce Department issues interim final rule restricting remote access to AI chips via overseas data centers11 sources
- Heat index 532Anthropic to begin trading on Nasdaq this week under ticker ANT9 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion

