Manifold Security
A security research firm focused on security vulnerability research for coding agents and developer toolchains.
Timeline (last 90 days)
Manifold Security disclosed that a repository's .git/config can, by setting core.fsmonitor, execute a helper program with the developer's full local privileges before any approval prompt is shown, outside the agent sandbox.
Manifold Security disclosed 8 vulnerabilities across 7 AI coding agents including Claude Code, Codex, and Cursor, which attackers can exploit through malicious configurations to induce the agents to execute programs.
All Facts (2)
Manifold Security disclosed that a repository's .git/config can, by setting core.fsmonitor, execute a helper program with the developer's full local privileges before any approval prompt is shown, outside the agent sandbox.
50%UnverifiedManifold Security disclosed 8 vulnerabilities across 7 AI coding agents including Claude Code, Codex, and Cursor, which attackers can exploit through malicious configurations to induce the agents to execute programs.
50%