Back to skill

Security audit

Voice Lead Recovery

Security checks for vulnerabilities and agentic risk

Overview

This skill clearly sends voicemail and lead details to BlueColumn for transcription, storage, recall, and follow-up, which matches its stated purpose but requires privacy care.

Before installing, confirm your business is allowed to upload call recordings, phone numbers, and lead notes to BlueColumn, and use it only for customers who have the required notice or consent. Avoid storing unnecessary sensitive details, manage the API key carefully, and confirm retention/deletion and access-control expectations with the service.

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
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (5)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill explicitly instructs users to transmit and store voicemail audio, phone numbers, service requests, urgency, objections, and financing-related details in a third-party service without any privacy warning, consent guidance, retention notice, or data-minimization advice. Because this is customer lead data and may include personal and potentially sensitive contextual information, the omission creates a real privacy and compliance risk rather than a harmless documentation issue.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
84% confidence
Finding

The documentation directs the agent to a third-party endpoint using a live API key pattern, confirming external data exfiltration from the local workflow to a vendor-controlled service. Although external API use is expected for SaaS functionality, it still represents a true security concern here because highly identifying caller data is being exported without trust-boundary warnings or verification guidance.

Content

Scanner excerpt · SKILL.md (reported line 19)May include surrounding context.

Send the voicemail or call audio; BlueColumn transcribes it and stores the structured lead.

bash
curl -X POST https://api.bluecolumn.ai/remember \
  -H "Authorization: Bearer $BLUECOLUMN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{

External Transmission

Medium
Category
Data Exfiltration
Confidence
84% confidence
Finding

The documentation directs the agent to a third-party endpoint using a live API key pattern, confirming external data exfiltration from the local workflow to a vendor-controlled service. Although external API use is expected for SaaS functionality, it still represents a true security concern here because highly identifying caller data is being exported without trust-boundary warnings or verification guidance.

Content

Scanner excerpt · SKILL.md (reported line 19)May include surrounding context.

Send the voicemail or call audio; BlueColumn transcribes it and stores the structured lead.

bash
curl -X POST https://api.bluecolumn.ai/remember \
  -H "Authorization: Bearer $BLUECOLUMN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{

External Transmission

Medium
Category
Data Exfiltration
Confidence
94% confidence
Finding

The recall query sends a caller's phone number and conversation history request to the external service, enabling aggregation and retrieval of cross-contact behavioral and promise-tracking data. This increases privacy risk because the skill encourages building a persistent memory profile tied to a directly identifying phone number, which can expose more customer history than is necessary for a single callback if mishandled or breached.

Content

Scanner excerpt · SKILL.md (reported line 34)May include surrounding context.

Pull the caller's history and open promises the moment you have the number.

bash
curl -X POST https://api.bluecolumn.ai/recall \
  -H "Authorization: Bearer $BLUECOLUMN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"q": "What do we know about Dana 602-555-0147 and what did we promise?"}'

External Transmission

Medium
Category
Data Exfiltration
Confidence
92% confidence
Finding

The note endpoint stores follow-up outcomes and communication preferences such as appointment timing and whether to text instead of call, which are personal operational details about an identified lead. In this skill's context, repeated storage of such notes creates a growing customer dossier in a third-party system without any mention of consent, retention limits, or controls over who can access or export it.

Content

Scanner excerpt · SKILL.md (reported line 45)May include surrounding context.

After the call, store the result so the next touch is informed.

bash
curl -X POST https://api.bluecolumn.ai/note \
  -H "Authorization: Bearer $BLUECOLUMN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text": "Dana booked inspection 09-30 10:00. Send financing PDF before visit. If not home, text instead of call.", "tags": ["lead", "roofing", "scheduled"]}'