Security Audit
Facebook Automation
github.com/ComposioHQ/awesome-claude-skillsTrust Assessment
Facebook Automation received a trust score of 82/100, placing it in the Mostly Trusted category. This skill has passed most security checks with only minor considerations noted.
SkillShield's automated analysis identified 2 findings: 0 critical, 1 high, 1 medium, and 0 low severity. Key findings include Agent access to Facebook Page Access Tokens, Potential Data Exfiltration via `file_url` in video uploads.
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 17, 2026 (commit 99e2a295). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings2
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| HIGH | Agent access to Facebook Page Access Tokens The `FACEBOOK_LIST_MANAGED_PAGES` tool explicitly returns `access_token` for each managed Facebook Page. These tokens grant significant control over the associated pages. While necessary for the agent to perform subsequent actions, direct exposure of these tokens to the agent's runtime environment or logs poses a high risk of credential harvesting and subsequent unauthorized access if the agent or its environment is compromised. Agents should be designed to handle these tokens securely, avoiding logging or persistent storage without strong encryption. Implement robust credential management practices within the agent. Ensure access tokens are not logged, are stored only in secure, encrypted vaults if persistence is required, and are used with the principle of least privilege. Consider token rotation and short-lived tokens where possible. The documentation should explicitly warn agent developers about the sensitive nature of these tokens. | LLM | SKILL.md:43 | |
| MEDIUM | Potential Data Exfiltration via `file_url` in video uploads The `FACEBOOK_CREATE_VIDEO_POST` tool accepts a `file_url` parameter for video uploads. If an attacker can control the input to this parameter, they could potentially provide a URL pointing to sensitive local files (e.g., `file:///etc/passwd`) or internal network resources. The underlying tool implementation might then fetch and upload this content to Facebook, leading to data exfiltration (Server-Side Request Forgery - SSRF). The documentation does not specify any URL validation or restriction mechanisms. The tool's backend implementation should strictly validate and sanitize `file_url` inputs. Restrict `file_url` to only allow specific, trusted domains or protocols (e.g., `https` only, disallow `file://`). Implement network access controls to prevent the tool from accessing internal resources based on external input. Agent developers should also be cautious about passing untrusted user input directly to this parameter. | LLM | SKILL.md:69 |
Scan History
Embed Code
[](https://skillshield.io/report/fae6e5781461d347)
Powered by SkillShield