Security Audit

canvas-automation

github.com/ComposioHQ/awesome-claude-skills

AI SkillCommit 27904475d127

CAUTION

Scanned about 2 months ago

Critical

Immediate action required

High

Priority fixes suggested

Medium

Best practices review

Low

Acknowledged / Tracked

Trust Assessment

canvas-automation received a trust score of 73/100, placing it in the Caution category. This skill has some security considerations that users should review before deployment.

SkillShield's automated analysis identified 2 findings: 0 critical, 2 high, 0 medium, and 0 low severity. Key findings include Dynamic Tool Execution via External MCP, Supply Chain Risk - Unpinned Dynamic Tool Dependencies.

The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. All layers scored 70 or above, reflecting consistent security practices.

Last analyzed on February 20, 2026 (commit 27904475). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.

Layer Breakdown

Manifest Analysis

100%

Static Code Analysis

100%

Dependency Graph

100%

LLM Behavioral Safety

70%

Behavioral Risk Signals

Network Access

2 findings

Shell Execution

2 findings

Dynamic Code

2 findings

Excessive Permissions

1 finding

Security Findings2

Severity	Finding	Layer	Location
HIGH	Dynamic Tool Execution via External MCP The skill delegates tool discovery and execution to the `rube` MCP via `RUBE_SEARCH_TOOLS`, `RUBE_MULTI_EXECUTE_TOOL`, and `RUBE_REMOTE_WORKBENCH`. This design means the agent implicitly trusts the `rube` MCP to provide and execute arbitrary tools, potentially granting broad access to external services (Canvas) and execution capabilities (`RUBE_REMOTE_WORKBENCH`). If the `rube` MCP or the tools it provides are compromised, the agent could be coerced into performing unintended or malicious actions. Implement strict allow-listing or sandboxing for tools provided by the MCP. Require explicit user confirmation for execution of newly discovered or high-risk tools. Implement robust input validation and output sanitization for all tool interactions. Consider limiting the scope of `RUBE_REMOTE_WORKBENCH` or requiring explicit permissions for its use.	LLM	SKILL.md:40
HIGH	Supply Chain Risk - Unpinned Dynamic Tool Dependencies The skill explicitly instructs the agent to dynamically discover tools using `RUBE_SEARCH_TOOLS` and warns against hardcoding tool slugs due to schema changes. This means the skill relies on unpinned, dynamically fetched tools from the `rube.app` MCP. There is no mechanism described to verify the integrity, version, or authenticity of these tools before execution. A compromise of the `rube.app` MCP could lead to the delivery and execution of malicious tools by the agent. Implement a mechanism to pin or hash tool definitions/schemas from the MCP. Require cryptographic signatures for tools provided by the MCP. Implement a trusted registry or allow-list for tool slugs and versions. Regularly audit the `rube.app` MCP for security vulnerabilities.	LLM	SKILL.md:50

Scan History

Embed Code

[![SkillShield](https://skillshield.io/api/v1/badge/44d05733e0d8eaaf.svg)](https://skillshield.io/report/44d05733e0d8eaaf)