Map / Runtime & Guardrails / PromptHalo
PromptHalo
prompthalo.ai ↗ · LinkedIn ↗ · Open on the map
Runtime & GuardrailsMCP & LLM GatewaysAI Red TeamingObservability & Governance
Capabilities
Model-agnostic platform that combines automated AI red-teaming (static probes, a prompt-injection library and PAIR adaptive attacks) with runtime enforcement over what AI agents are allowed to do. It also offers an agent registry with per-agent policies, multi-agent workflow topology and risk views, behavioral baselines, and audit trails of AI decisions.
Sensors / integration
Deployed as a security sidecar next to production AI applications, as an API gateway in front of LLM endpoints (OpenAI, Anthropic, Ollama, vLLM, LiteLLM and generic HTTP+JSON), through agent-framework/orchestration integrations, and with a browser extension that captures chatbot targets for scanning.
Protections
Inspects prompts and agent actions inline and blocks or flags prompt injection, retrieval poisoning, PII exfiltration and unauthorized tool or API calls before they execute. Each agent gets an identity key, passports and scoped tool permissions that can be revoked.
Differentiator
Works without access to the model, its training data or architectural changes. It focuses on controlling AI-triggered actions before execution, and feeds red-team findings directly into runtime enforcement policies.
Team & funding
- Founders
- Madhuri Chandoor
Other Runtime & Guardrails companies
AariaSec · ActaClad · ActiveFence · Acuvity · AGAT Software · AgentFox · AI Aramid · AI Prompt Shield · Aiceberg · Aim Security · AIR · Aira Security · See all →
Last updated · Get the “Listed on” badge