Security audit

OpenCawl

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent phone-calling integration, but users should treat call transcripts, recordings, and contact details as sensitive.

Install only if you trust OpenCawl with phone numbers, call goals, context, transcripts, recordings, voicemails, and extracted contact data. Confirm consent and recording-law obligations before making or receiving calls, and configure inbound webhooks or gateway tokens only for endpoints you control.

A.I.G

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

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (8)

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill explicitly handles phone calls and its documented status output includes sensitive artifacts such as call summaries, extracted contact details, transcripts, and recording URLs. While it mentions HTTPS requests and an API key, it does not clearly warn users that personal contact data and conversation content may be transmitted to OpenCawl and retained as transcripts/recordings, which creates a meaningful privacy and consent risk.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The API supports placing calls and retrieving transcripts, recordings, summaries, phone numbers, and extracted personal data, yet the documentation provides no warning about consent, privacy, or handling of sensitive communications data. In a telephony context this is dangerous because misuse can violate call-recording laws, expose PII, and encourage unsafe storage or sharing of highly sensitive conversation artifacts.
Content: No source excerpt is available for this finding.

External Transmission

Medium

Category: Data Exfiltration
Confidence: 50% confidence
Finding: Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
Content: Scanner excerpt · SKILL.md (reported line 90)May include surrounding context.
md
"availability": ["Thursday 2pm ET", "Friday 10am ET"] }, "transcript": "[{\"role\":\"agent\",\"text\":\"Hi, may I speak with Jamie?\"},...]", "recording_url": "https://api.twilio.com/...", "created_at": "2026-03-28T14:00:00Z", "completed_at": "2026-03-28T14:03:07Z" }

External Transmission

Medium

Category: Data Exfiltration
Confidence: 50% confidence
Finding: Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
Content: Scanner excerpt · api.md (reported line 70)May include surrounding context.
md
"availability": ["Thursday 2pm ET", "Friday 10am ET"] }, "transcript": "[{\"role\":\"agent\",\"text\":\"Hi, may I speak with Jamie?\"},...]", "recording_url": "https://api.twilio.com/...", "created_at": "2026-03-28T14:00:00Z", "completed_at": "2026-03-28T14:03:07Z" }

External Transmission

Medium

Category: Data Exfiltration
Confidence: 50% confidence
Finding: Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
Content: Scanner excerpt · api.md (reported line 231)May include surrounding context.
md
"availability": ["Thursday 2pm ET", "Friday 10am ET"] }, "transcript": "[{\"role\":\"agent\",\"text\":\"Hi, may I speak with Jamie?\"},...]", "recording_url": "https://api.twilio.com/...", "created_at": "2026-03-28T14:00:00Z", "completed_at": "2026-03-28T14:03:07Z" }

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The documented API surface includes inbound/autonomous call handling and task dispatch features that go beyond the manifest’s narrower description. This creates a scope-mismatch risk: an agent or integrator may grant or invoke capabilities they did not expect, enabling unreviewed call automation, webhook/task dispatch, or inbound processing paths with different security and privacy implications.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: Phone number provisioning and inbound configuration are materially different powers than simply placing outbound calls and checking outcomes. If these capabilities are exposed without being reflected in the manifest, users may unknowingly allow number assignment, inbound routing changes, prompt changes, or gateway webhook configuration that alters account behavior and can redirect sensitive call flows.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The advanced task polling and completion endpoints introduce an additional workflow for inbound task handling that is not described in the manifest. Hidden workflow capabilities can let an agent interact with dispatched inbound tasks, mark statuses, and submit result payloads without users realizing the skill can participate in autonomous inbound orchestration.
Content: No source excerpt is available for this finding.