T02 · Agent Memory Poisoning
Warning
- Location
SKILL.md:29- Finding
Unvalidated Conversation Content Can Be Persisted into Long-Term Agent Memory
- Content
View full analysis
- Remediation
View remediation
Security audit
Semantic Memory Boost
Security checks for vulnerabilities and agentic risk
Overview
The skill is a coherent memory helper, but it tells the agent to read long-term memory and automatically persist conversation insights without clear consent, scope, or deletion controls.
Install only if you are comfortable with the agent using and updating memory. Before use, require explicit approval for each memory write, keep memory files project-scoped, and avoid storing credentials, private personal data, confidential business details, or instructions that could change future agent behavior.
Vulnerability Patterns
- Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
- Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
- 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
Findings (1)
Vulnerability Patterns
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- 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
Findings (2)
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 88% confidence
- Finding
The skill content is primarily written in Chinese and does not document a locale requirement or offer a language choice, which can cause operators or users to misunderstand memory-handling behavior, constraints, or failure modes. In a skill that performs retrieval and persistent memory actions, such ambiguity can indirectly increase security and privacy risk by causing incorrect use or uninformed consent.
- Content
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 94% confidence
- Finding
The skill explicitly instructs the agent to persist conversation-derived insights into a memory file after answering, but it does not warn the user or require consent before storing potentially sensitive or proprietary information. This creates a real privacy and data-governance risk because users may disclose confidential project details without realizing they will be retained beyond the current session.
- Content