Context Window Credential Leak

Name: AI Agent Threat Matrix
Creator: OpenA2A
License: https://www.apache.org/licenses/LICENSE-2.0

Exploit context window contents to leak credentials that were previously processed by the agent

Tactic

Credential Harvest · Stage 3

Extract API keys, tokens, and credentials from agent context and connected services

Attack class

RETROACTIVE-PRIV

Exploiting previously granted access or cached credentials to gain unauthorized capabilities

Evidence grade

Validated

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

DVAA validation

LegacyBot credential leak

Reproductions in Damn Vulnerable AI Agent, the OpenA2A intentionally-broken agent for kill-chain validation.

Honeypot

AgentPwn coverage

Live

context-windowagentpwn.com/learn ↗

An AgentPwn trap page produces a payload tagged with this technique class. Following the AgentPwn taxonomy of trap pages shows what an agent encounters.

Context-window tiers leak credentials retained in the active window.

Detect

Detection · HackMyAgent

Live2 live · 0 queued

CRED-001AGENT-CRED-001

npx hackmyagent secure --ciLive = implemented in hackmyagent; queued = declared

Defend

Defense · OASB controls

Live5 live · 0 queued

OASB 5.1 OASB 5.2 OASB 5.3 OASB 5.4 OASB 5.5

Live = documented at oasb.ai; queued = declared

Reference

How to cite

AI Agent Threat Matrix T-3006 (Context Window Credential Leak). OpenA2A, 2026. https://threats.opena2a.org/techniques/T-3006

← Back to the matrix