Back to skill

Security audit

小星记忆系统

Security checks for vulnerabilities and agentic risk

Overview

This is a persistent memory skill that fits its stated purpose but gives the agent broad automatic access to private and long-term memory without clear user consent or retention controls.

Install only if you want the agent to keep and search long-term files about you and prior work. Before use, define what may be stored, exclude secrets and sensitive personal data, and require explicit approval before reading private memory or saving psychological or profile-like notes.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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
Findings (1)

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
references/architecture.md:51
Finding

Automatic and Overbroad Access to Private Agent Memory

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill description is very broad, covering nearly any situation involving history, preferences, names, dates, or cross-session information, without clear invocation limits or exclusion criteria. In an agent setting, this can cause over-triggering and unnecessary access to persistent memory, increasing the chance that unrelated or sensitive user data is retrieved or stored when not needed.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill enables persistent storage of cross-session memory but does not warn users that their information may be written to durable files, including potentially private or sensitive data. Because the workflow explicitly encourages saving history, preferences, dates, and personal details, the absence of consent and sensitivity guidance materially increases the risk of privacy violations, over-collection, and accidental retention of secrets.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The document explicitly instructs the agent to read a user's private memory file during session startup, but it provides no consent boundary, purpose limitation, or warning about handling sensitive personal data. In a persistent-memory skill, automatic loading of private memory increases the chance of over-collection, unintended disclosure in downstream prompts, and use of sensitive data outside the user's expectations.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The instruction that any 'psychological note' must be written to a file encourages persistence of highly sensitive personal inferences without notice, minimization, or retention controls. In a memory system designed for cross-session recall, this can create lasting sensitive profiles about a user that may later be surfaced, misused, or exposed if the files are accessed by other components or operators.

Content

No source excerpt is available for this finding.