Security audit

Clawcall Free

Security checks across malware telemetry and agentic risk

Overview

The skill is a coherent outbound-calling integration, but it needs Review because it can place real phone calls, save credentials and phone numbers persistently, and trigger broad research or local diagnostics without clear user confirmation boundaries.

Review before installing. Use this only if you are comfortable with an agent placing real outbound calls, sending the call task content to an external voice-call API, storing an API key and phone number across sessions, and potentially doing public business lookups. Ask for explicit confirmation before each call and avoid granting broad exec access unless local network diagnostics are truly needed.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Context-Inappropriate Capability

Medium

Confidence: 92% confidence
Finding: The workflow explicitly tells the agent to perform independent public/business-information reconnaissance before placing calls, which expands behavior beyond the declared phone-calling integration. In an agent environment with broad tool access, this can trigger unbounded browsing, scraping, or external data collection without clear user consent or scope limits.

Context-Inappropriate Capability

Medium

Confidence: 97% confidence
Finding: The error-handling section instructs the agent to run ping/network diagnostics, which is unrelated to the stated calling API functionality and encourages use of host-system execution capabilities. In environments exposing exec, this broadens the skill into system probing and could be abused to run unauthorized commands or reveal network details.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The capability-coverage section is overly broad and ambiguous, listing many fragmented phrases as covered scenarios rather than narrowly scoped actions. This can cause accidental activation for loosely related requests and increase the chance the agent applies the skill in contexts the user did not intend, including real-world calling and data handling.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The skill stores API keys and user phone numbers in persistent local state but does not provide a clear user-facing warning or consent mechanism. Persisting credentials and PII across sessions increases the risk of unintended reuse, disclosure to other workflows, or compromise if local files are accessible.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The description explains the feature set but does not prominently warn that using the skill will place real outbound phone calls to external parties. Because this creates real-world side effects, costs, and privacy implications, lack of explicit warning increases the chance of accidental or non-consensual activation.

VirusTotal

VirusTotal findings are pending for this skill version.

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: SKILL.md:116