AI Agent Threat Matrix
Matrix/Initial Access/T-2005
T-2005validated

Tool Description Injection

Embed malicious instructions in tool or skill descriptions that execute when the agent processes them

Tactic

Initial Access (Stage 2)

Gain control over agent behavior through prompt manipulation or input exploitation

Attack Class

SKILL-FRONTMATTER

Embedding malicious instructions in skill or plugin metadata and description fields

Evidence

validated

Reproduced in controlled lab environment (DVAA) with documented steps.

DVAA Validation

PluginBot

Detection (HackMyAgent)

SKILL-001SKILL-004SKILL-005
npx hackmyagent secure --ci

Defense (OASB Controls)

OASB 3.1OASB 3.2OASB 3.3OASB 3.4OASB 3.5

How to Cite

AI Agent Threat Matrix T-2005 (Tool Description Injection). OpenA2A, 2026. https://threats.opena2a.org/techniques/T-2005
← Back to Matrix