What Happened
This report covers two recent AI-related security incidents: a supply-chain attack on Mercor linked to the LiteLLM open-source project, and a separate leak of Anthropic source code attributed to human error.[4] The Mercor case involved alleged access to client data claimed by the Lapsus$ group, underscoring AI supply chain and third-party library risks for startups and fintech platforms using LLM infrastructure.[4]
Why It Matters
The report describes two AI security incidents: a supply-chain compromise affecting Mercor through the open-source LiteLLM ecosystem, and a separate source-code leak at Anthropic attributed to human error. The Mercor case highlights how third-party AI infrastructure and dependencies can expose sensitive client and operational data, while the Anthropic incident shows that ordinary data-handling mistakes can still create material risk. RealGround should treat this as strong evidence that AI startups and fintech-style platforms need dependency inventorying, artifact verification, access controls, and incident-ready review of third-party AI components.
RealGround Analysis
This signal maps to AI supply chain. Organizations using AI agents, LLM APIs, SaaS integrations, or sensitive data workflows should review whether this class of issue could create unauthorized tool execution, data leakage, weak approval gates, or unmanaged supply-chain exposure.
Recommended Actions
- Restrict AI agent tool permissions and production write paths.
- Review sensitive data access across prompts, logs, embeddings, memory, and SaaS integrations.
- Add human approval workflows for high-impact or state-changing actions.
- Run prompt injection and indirect prompt injection tests against affected workflows.
- Document the owner, control gap, and remediation deadline for this risk class.