Whitespace.
See who already built your idea — and the space they left open.

CLUSTER ADVERSARIAL-AGAINST-AGENTS · 19 ENTRIES · EARLIEST ENTRY 2026-08-18

AI agent adversarial testing

Platforms that test AI agents against adversarial attacks in sandboxed or synthetic environments, exposing failure modes and hardening tool permissions. The cluster's dated history is too thin to read a shape from—only two dated entries exist before the index start, so it predates the index or its members are undated. Traction leaders are Kintsugi_Demo and Agent-security (3 stars each), with guardrail-auditor at 1 star.

Entry timeline

ENTRY TIMELINE

JAN 2026: 1APR 2026: 1MAY 2026: 1AUG 2026: 3SEP 2026: 10OCT 2026: 3
JAN 2026OCT 2026

Each bar is the month a member carries, which is a creation date where the source published one and a crawl date otherwise. Nothing in the index was found before 11 August 2026, and coverage broadened sharply on 14–15 August, so a bar from then on reports how much we found that month, not how much was built; an earlier bar reports how far back a member's own source dated it — not how long we have been watching.

Members by activity

Activity signals, not revenue.

Build origin · 2 hand-coded · 17 unlabelled

Operator · 8 hobby · 3 indie · 1 commercial · 7 unlabelled

Two labels per entry, never collapsed into one: how it was built, and who stands behind it. Counted over the 19 listed below. Unlabelled means nothing has decided yet — not that the answer is small.

  1. 01guardrail-auditor1★26 daysliveindieWeb platform that runs adversarial security tests against AI chat systems to detect prompt injection, data leakage, and role bypass, then scores risk.
  2. 02Kintsugi_Demo3★17 days—Web platform that tests AI agents against adversarial attacks using a policy-enforced tool gateway in a synthetic environment.
  3. 03discoverylab1★14 days—hobbyMulti-agent research application that generates testable scientific hypotheses from literature and employs an adversarial critic to actively try to disprove them.
  4. 04Agent-security3★1 mo—Runs structured adversarial attacks against AI agents to test security, inspect evidence, and harden tool permissions.
  5. 05agentic-security-lab0★8 dayslivehobbyInteractive security lab that lets users attack and defend deliberately vulnerable AI agents in a browser-based cyber range.
  6. 06AI-Agent-Security-Red-Team-Platform0★14 days—Security platform that autonomously attacks AI agent systems, scores vulnerabilities, and gates regressions via a dashboard and CI.
  7. 07pinned-not-checked0★22 dayslivehobbyInteractive simulator that demonstrates how AI coding agents can install tampered plugins while SHA-pin checks falsely report success.
  8. 08Agent-Plus-Project0★23 days—Web platform for evaluating AI agents by running test cases and measuring reliability and hallucination rates.
  9. 09CyberGuard0★18 dayslivehobbySecurity tool that investigates, verifies, and governs AI agent actions through evidence-based analysis and independent review.
  10. 10trustmesh0★5 daysliveWeb platform that tests autonomous AI agents against adversarial scenarios, scores their safety risks, and recommends guardrails before deployment.
  11. 11AgentGuard2★1 mo—hobbySecurity proxy that intercepts and guards AI agent tool calls with policy rules and human approval.
  12. 12fulcrum0★26 days—hand-codedindieA set of harness-agnostic agent skills for long unattended OutSystems ODC builds, encoding field-derived traps, verification ladders, and guardrails.
  13. 13Flash0ver1★29 days—Runtime security platform that monitors autonomous AI agent swarms, reconstructs incidents, and contains dangerous capabilities as they emerge.
  14. 14AgentGuard0★1 molivehobbyPlatform that tests AI agents against adversarial scenarios in a sandbox to expose failure modes and measure reliability before release.
  15. 15TrustLayer--Security-Control-Plane-For-LLM-AI5★8 mostaleliveindieSecurity service that firewalls AI agents and LLM applications against prompt injection, tool hijacking, and behavioral drift in production.
  16. 16ai-agent-launch-tools0★4 molivehand-codedhobbySet of dependency-free tools and checklists for securing AI agents and MCP servers before launch.
  17. 17thinking&thoughts - LLM Research Platform0★unknownliveResearch platform for documenting and sharing adversarial analyses of large language models and AI safety.
  18. 18MIR[AI]GE0★unknownlivehobbyWeb app that deploys a generative decoy to lure AI-augmented attackers and waste their computing resources.
  19. 19Realm Labs | We Stop AI Failures0★unknownlivecommercialPlatform that monitors AI application behavior at runtime and flags reasoning failures that evaluation tools, gateways, and guardrails miss.

Ranked by traction, highest first, and paged 20 at a time — this is a window on the cluster, never the whole of it. Ties are broken by a stable id so a product cannot appear on two pages or fall between them. Browse every cluster.