Security audit

bbc-skill-tool

Security checks for vulnerabilities and agentic risk

Overview

This is an instruction-only BuilderBot Cloud WhatsApp bot skill with privacy cautions around CRM/webhook examples, but no hidden code or deceptive behavior.

Install only if you intend to manage BuilderBot Cloud WhatsApp bots. Before using the CRM, webhook, scraping, or handoff patterns in production, add user-facing privacy notices, collect only necessary data, use trusted endpoints, protect credentials, redact sensitive conversation content, and require confirmation before deployment or deletion actions.

SkillSpector

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 (4)

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The example explicitly collects personal data such as name and email and sends it to an external CRM/API, but provides no user notice, consent step, minimization guidance, or security constraints. In a bot-building skill, this pattern is likely to be copied directly into production flows, creating real privacy, compliance, and data-handling risk.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: Posting conversation context to an agent dashboard can expose sensitive user messages, contact details, or other confidential content if done without notice or filtering. The skill presents this as a standard handoff pattern, which normalizes broad sharing of chat transcripts without discussing consent, redaction, or least-privilege access.

Ssd 3

Medium

Confidence: 96% confidence
Finding: This flow operationalizes collection of user contact data followed by transmission to an external CRM, making the privacy issue concrete rather than hypothetical. Because the skill is instructional and aimed at real business bots, the omission of safeguards increases the chance that operators will implement personal-data collection in an unsafe or noncompliant manner.

Ssd 3

Medium

Confidence: 93% confidence
Finding: Sending full conversation context during handoff may leak sensitive content beyond what is needed for support, especially if the downstream dashboard is less protected or broadly accessible. The pattern is more dangerous in this skill context because it is presented as an advanced best-practice-style recipe for deployment in customer-facing WhatsApp bots.

Static analysis

No suspicious patterns detected.