- ID
- AIA-004
- Category
- AI AI agent incident
- Disclosed
- 2025-05
- Systems
- GitHub MCP server
- Harm type
- Data leak / prompt injection
- Status
- reported
Reported by the source; no vendor confirmation recorded in seed data.
Summary
Invariant Labs showed that a malicious public GitHub issue could steer an agent using the GitHub MCP server into putting private-repo contents into a public PR.
Controls
Controls that could prevent, detect, or support recovery from this kind of failure. Mappings are editorial judgments made by the method, not findings about what any party should have done.
NIST AI RMF
- MAP 4.1 Risks from third-party components are mapped
- MANAGE 2.4 Mechanisms exist to supersede, disengage, or deactivate the system
NIST SP 800-53
Human-approval (JCR) checkpoint
Sources
- invariantlabs.ai primary link checked 2026-10-31, HTTP 200
Post dated 2025-05-26
https://invariantlabs.ai/blog/mcp-github-vulnerability
Corrections
- Verification round 1: date month and summary verified against the Invariant Labs post (2025-05-26).
To report an error in this entry, see corrections.