Back to skill

Security audit

JSON画布工具

Security checks for vulnerabilities and agentic risk

Overview

This JSON Canvas skill is not clearly malicious, but it asks for command execution and API-related authority that is broader than its stated canvas-editing purpose.

Review before installing. Use this only if you are comfortable granting a Markdown skill command-execution capability and vague API/key handling for what appears to be a local JSON Canvas task. Prefer a version that removes exec, narrows the description to canvas files, and documents exact inputs, outputs, network use, and credential handling.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (8)

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

Documenting command execution for a skill whose stated purpose is editing JSON Canvas files is a significant risk escalation. In an agent setting, unnecessary exec capability can be abused to run arbitrary or unsafe shell commands, access local data, modify the environment, or pivot beyond the intended file-editing scope.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill is presented as a JSON Canvas editor but also advertises API integration, webhook/system connection, and broader automation behavior unrelated to local canvas editing. This scope expansion can cause an agent or user to authorize networked or higher-risk actions under a misleadingly narrow name, increasing the chance of unintended data exposure or misuse.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The invocation guidance is overly broad and ambiguous, spanning file handling, format conversion, content extraction, API integration, webhooks, and system connection tasks. Ambiguous trigger conditions increase the likelihood that the skill is selected for inappropriate contexts, where its tools or implied permissions may be excessive for the task.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The core functionality summary uses vague, trigger-style language without clearly defining operational scope or boundaries. While less severe than explicit exec or network claims, this ambiguity still increases misrouting risk and reduces a user's ability to judge whether the skill is safe for a given task.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The documentation says the skill provides JSON Canvas creation/editing, but the feature table states that capability is unavailable in the free version. Such contradictions undermine trust and can mislead operators about what actions the skill will actually attempt, which is risky in agentic environments where users depend on accurate capability boundaries.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The documented response format is a grading/report schema rather than a JSON Canvas file or edit result. This mismatch suggests the skill may return data unrelated to its stated purpose, which can confuse downstream automation and cause unsafe assumptions about output handling or validation.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill introduces API key setup and implied external API/network use even though its stated purpose is local JSON Canvas file editing. Unnecessary credential collection and network dependency expand the attack surface and create opportunities for secret exposure, overpermissioning, or unexpected outbound data flows.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill claims capabilities like data synchronization, multi-source queries, chart generation, and external system integration that materially exceed simple canvas editing. This kind of capability inflation is dangerous because it normalizes broader access and may induce the agent to perform file, network, or data operations outside the user's reasonable expectations.

Content

No source excerpt is available for this finding.