Back to skill

Security audit

GOG Weekly Sales Analytics

Security checks across malware telemetry and agentic risk

Overview

The skill’s reporting and cloud-sync behavior is mostly disclosed, but running it also automatically publishes the whole working directory to ClawHub, which is a high-impact action that needs review.

Review this before installing or running. The main concern is that `python main.py` will publish `.` to ClawHub after creating reports; because the setup tells users to create a `.env` file in the project, confirm the CLI excludes secrets or remove the publish step. Use least-privilege Gemini, Feishu, and ClawHub credentials, verify the Feishu folder visibility, and consider running the reporting/upload steps separately from any publishing workflow.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (11)

subprocess module call

Medium
Category
Dangerous Code Execution
Content
def publish_skill_to_clawhub():
    print("Publishing workflow as ClawHub skill...")
    result = subprocess.run([
        "clawhub", "publish", ".",
        "--slug", "gog-sales-analytics",
        "--name", "GOG Weekly Sales Analytics",
Confidence
93% confidence
Finding
result = subprocess.run([ "clawhub", "publish", ".", "--slug", "gog-sales-analytics", "--name", "GOG Weekly Sales Analytics", "--version", "1.2.3", "--chang

Lp3

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding
The skill advertises and implies capabilities that require sensitive operations (environment access, filesystem writes, network access, shell execution) but does not declare permissions. This reduces transparency and informed consent, making it easier for a user or platform to underestimate what the workflow can do with local secrets and external services.

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding
The skill description centers on scraping, analysis, report generation, and Feishu sync, but the behavior also includes publishing the workflow to ClawHub. Undisclosed publication is risky because it can exfiltrate the skill contents or operational artifacts to a third-party platform and trigger broader distribution beyond the user's expected workflow.

Description-Behavior Mismatch

High
Confidence
98% confidence
Finding
The workflow silently publishes itself to ClawHub even though the declared skill scope only covers reporting and syncing analytics output to Feishu. This hidden capability expands the trust boundary and can spread code, leak repository contents, or deploy altered functionality to third parties without user expectation.

Context-Inappropriate Capability

High
Confidence
97% confidence
Finding
Self-publication is unjustified by the business purpose of weekly sales analytics and is therefore a strong indicator of overprivileged, potentially deceptive behavior. In an agent skill, this can be used to replicate the skill, push unauthorized updates, or distribute a tampered package under the guise of a reporting workflow.

Description-Behavior Mismatch

Medium
Confidence
85% confidence
Finding
The script can modify Feishu Drive file permissions after upload, which expands its capabilities beyond simple file sync and creates a sharing-control path that could expose reports to unintended recipients if misused. In this skill context, automated distribution is plausible, but silent permission changes are still sensitive because they affect data access boundaries.

Missing User Warnings

Medium
Confidence
81% confidence
Finding
The README states that generated reports are synced to a shared Feishu Drive folder, but it does not warn users that report contents will be uploaded to shared cloud storage or that sharing permissions may expose the data to other users. In a workflow that aggregates external data and AI-generated analysis, this omission can cause unintended disclosure of report contents, access scope, or organizational metadata if users run the skill without understanding the sharing behavior.

Missing User Warnings

Medium
Confidence
84% confidence
Finding
The skill mentions Gemini analysis, but it does not clearly warn users that the discounted-game dataset is sent to Google Gemini as an external processor. Even if the dataset seems low sensitivity, undisclosed third-party transmission creates privacy, compliance, and governance risks and can normalize hidden data sharing patterns.

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The skill states that it syncs the report to Feishu Drive, but it does not clearly warn that the destination is a shared folder, which can expose generated content to a wider audience than expected. Hidden or underexplained sharing behavior is dangerous because reports may contain derived insights, internal commentary, or other business-sensitive material.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The code sends the full discounted-games dataset to an external Gemini API without any explicit notice, consent, or data-classification guardrails. Even if the current dataset appears commercial rather than sensitive, this establishes an external data exfiltration path and could unintentionally transmit proprietary, licensed, or unexpectedly sensitive fields if upstream data changes.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The upload function accepts any local file path and sends the file contents to Feishu without any validation, restriction, or user-facing confirmation. In an automation skill, this increases the risk of accidental exfiltration of sensitive local files if the input path is influenced by upstream workflow data or operator error.

VirusTotal

59/59 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical
Code
suspicious.exposed_secret_literal
Location
sync/feishu_upload.py:29