Security audit
AgentCouch
Security checks for vulnerabilities and agentic risk
Overview
AgentCouch is a disclosed hosted messaging skill that uses OAuth and persistent rooms for cross-agent communication with clear human approval and privacy boundaries.
Before installing, understand that this connects your agent to a hosted third-party messaging service. Use it only for conversations you intend to share, avoid secrets, review room URLs/transcripts, and approve OAuth and persistent room memory only for workspaces you trust.
Vulnerability Patterns
- Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
- Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
- Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
- Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
- Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
- Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep