Back to skill

Security audit

Sumo Smart Note

Security checks for vulnerabilities and agentic risk

Overview

This note-taking skill does what it says, but it automatically copies development notes and conversation content into an outside shared notebook path without a separate confirmation step.

Install only if you intend notes to be saved both in workspace memory and in the SumoNoteBook raw/shared folder. Avoid using it for secrets, credentials, proprietary discussions, or private conversation text unless you are comfortable with that shared location retaining the content.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (10)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding

The skill describes file read/write behavior and access to environment-backed local paths but does not declare any explicit tool scope or permissions. This creates a mismatch between documented behavior and authorization boundaries, making it easier for an agent runtime to grant broader-than-necessary capabilities or for reviewers to miss that the skill writes to local and shared storage.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The description says the skill automatically saves notes to both the workspace and SumoNoteBook, but it does not clearly warn users that one destination is a shared path. This weak transparency increases the risk that users disclose development details, secrets, or private conversation content without understanding that the data will be duplicated into broader-access storage.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The trigger phrases are common conversational expressions like '記下來' and '寫下來', which can easily appear in ordinary dialogue without the user intending a persistent write or sync. Because this skill stores content in multiple locations, broad triggers raise the chance of accidental activation and unintended recording of sensitive information.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

This section explicitly instructs the agent to record conversation content, development process, problems, and solutions, then sync them into a shared storage path. In context, that creates a natural-language exfiltration channel where secrets, internal code details, credentials, or proprietary discussion may be persistently copied to a broader audience or less-controlled repository.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The example normalizes automatic copying of user statements into shared storage using an ordinary note-taking request. That makes the risky behavior feel routine and lowers the chance that users recognize they are publishing potentially sensitive material beyond the local workspace.

Content

No source excerpt is available for this finding.

Tainted flow: 'filepath' from os.environ.get (line 45, credential/environment) → open (file write)

Medium
Category
Data Flow
Confidence
65% confidence
Finding

Data from a source is assigned to a variable that is later passed to a sink, creating a variable-mediated taint flow.

Content

Scanner excerpt · smart_note.py (reported line 62)May include surrounding context.

python
else:
        new_content = content
    
    with open(filepath, 'w', encoding='utf-8') as f:
        f.write(new_content)
    
    print(f"[OK] Saved to workspace: {filepath}")

Tainted flow: 'filepath' from os.environ.get (line 45, credential/environment) → open (file write)

Medium
Category
Data Flow
Confidence
65% confidence
Finding

Data from a source is assigned to a variable that is later passed to a sink, creating a variable-mediated taint flow.

Content

Scanner excerpt · smart_note.py (reported line 101)May include surrounding context.

python
else:
        new_content = content
    
    with open(filepath, 'w', encoding='utf-8') as f:
        f.write(new_content)
    
    print(f"[OK] Saved to workspace: {filepath}")

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill automatically duplicates note content into a second external shared directory with no confirmation, warning, or opt-in control. In a note-taking context, users may record secrets, internal code, credentials, or sensitive project details; silent cross-sync expands data exposure and can leak information beyond the workspace's expected boundary.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The description and trigger examples are entirely specified in Chinese, including the expected invocation phrase, without indicating any language choice or opt-in. Under the policy, forcing a specific language or locale without user selection can be a natural-language policy issue.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
96% confidence
Finding

SQP-3 applies to all file types and covers language or locale policy violations. The file's natural-language interface, usage text, and user-facing messages are written in Chinese only, with no opt-in or explanation that the skill is intended exclusively for a Chinese-language environment.

Content

No source excerpt is available for this finding.