Forwarded from AISecHub
Agent Skills for Large Language Models: Architecture, Acquisition, Security, and Path Forward
As agents load “skills” (instructions/code/resources) on demand—e.g., via MCP—the trust boundary shifts from model weights to the skill acquisition/runtime layer. The practical security question is how progressive disclosure and portable skill definitions constrain untrusted skills so they can’t silently expand tool privileges or data access.
#MCP #AgentSecurity #AISecurity #Research
https://openreview.net/forum?id=Er0p92BsmW
As agents load “skills” (instructions/code/resources) on demand—e.g., via MCP—the trust boundary shifts from model weights to the skill acquisition/runtime layer. The practical security question is how progressive disclosure and portable skill definitions constrain untrusted skills so they can’t silently expand tool privileges or data access.
#MCP #AgentSecurity #AISecurity #Research
https://openreview.net/forum?id=Er0p92BsmW
openreview.net
Agent Skills for Large Language Models: Architecture, Acquisition...
The transition from monolithic language models to modular, skill-equipped agents marks a defining shift in how large language models (LLMs) are deployed in practice.
Rather than encoding all...
Rather than encoding all...
Forwarded from AISecHub
AI Coding Agents as Attack Surface: MCP, Poisoning, and Miasma
The core risk is a trust-boundary mismatch: coding agents ingest untrusted inputs (e.g., MCP tool outputs or repo content) but can act with developer-level privileges, so poisoning and instruction smuggling can turn “context” into autonomous actions.
#MCP #AgentSecurity #AISecurity #Report
https://labs.cloudsecurityalliance.org/research/ai-coding-agent-attack-surface-mcp-repository-poisoning-v1-0
The core risk is a trust-boundary mismatch: coding agents ingest untrusted inputs (e.g., MCP tool outputs or repo content) but can act with developer-level privileges, so poisoning and instruction smuggling can turn “context” into autonomous actions.
#MCP #AgentSecurity #AISecurity #Report
https://labs.cloudsecurityalliance.org/research/ai-coding-agent-attack-surface-mcp-repository-poisoning-v1-0
Lab Space
AI Coding Agents as Attack Surface: MCP, Poisoning, and Miasma
AI Coding Agents as Attack Surface: MCP, Poisoning, and Miasma Executive Summary The rapid adoption of AI coding agents has introduced a category of security risk the industry has not previously en…