Documentation
¶
Overview ¶
AgentExploitation implements attacks against AI coding editor agents using the AIShellJack pattern: 75-88% execution rate, 71.5% privilege escalation.
Source: IEEE S&P 2026 (arXiv 2511.05797)
Package tool_use implements attacks targeting AI tool/function calling interfaces.
These attacks exploit the function-calling / tool-use interface that modern LLMs provide, transforming harmful queries into seemingly legitimate tool invocations.
Index ¶
- type AgentExploitationModule
- func (m *AgentExploitationModule) Category() common.AttackCategory
- func (m *AgentExploitationModule) Description() string
- func (m *AgentExploitationModule) Execute(ctx context.Context, provider common.Provider, config common.AttackConfig) (*common.AttackResult, error)
- func (m *AgentExploitationModule) Name() string
- func (m *AgentExploitationModule) Techniques() []common.TechniqueInfo
- type IMISTModule
- func (m *IMISTModule) Category() common.AttackCategory
- func (m *IMISTModule) Description() string
- func (m *IMISTModule) Execute(ctx context.Context, provider common.Provider, config common.AttackConfig) (*common.AttackResult, error)
- func (m *IMISTModule) Name() string
- func (m *IMISTModule) Techniques() []common.TechniqueInfo
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type AgentExploitationModule ¶
type AgentExploitationModule struct{}
AgentExploitationModule implements AI coding editor exploitation attacks.
Source: IEEE S&P 2026 (arXiv 2511.05797)
func (*AgentExploitationModule) Category ¶
func (m *AgentExploitationModule) Category() common.AttackCategory
func (*AgentExploitationModule) Description ¶
func (m *AgentExploitationModule) Description() string
func (*AgentExploitationModule) Execute ¶
func (m *AgentExploitationModule) Execute(ctx context.Context, provider common.Provider, config common.AttackConfig) (*common.AttackResult, error)
func (*AgentExploitationModule) Name ¶
func (m *AgentExploitationModule) Name() string
func (*AgentExploitationModule) Techniques ¶
func (m *AgentExploitationModule) Techniques() []common.TechniqueInfo
type IMISTModule ¶
type IMISTModule struct{}
IMISTModule transforms harmful queries into tool invocations to bypass safety filters that only check natural language content.
Source: arXiv 2601.05466
func (*IMISTModule) Category ¶
func (m *IMISTModule) Category() common.AttackCategory
func (*IMISTModule) Description ¶
func (m *IMISTModule) Description() string
func (*IMISTModule) Execute ¶
func (m *IMISTModule) Execute(ctx context.Context, provider common.Provider, config common.AttackConfig) (*common.AttackResult, error)
func (*IMISTModule) Name ¶
func (m *IMISTModule) Name() string
func (*IMISTModule) Techniques ¶
func (m *IMISTModule) Techniques() []common.TechniqueInfo
Source Files