Back to skill

Security audit

openclaw-whobot-skill

Security checks for vulnerabilities and agentic risk

Overview

This is mostly a WhoBot knowledge-base skill, but it should be reviewed because it promotes human-like AI calling and customer-data workflows without enough disclosure, consent, or network-hardening guidance.

Review this carefully before installing if you will rely on it for customer-facing, healthcare, telecom, sales, or support workflows. Require clear AI disclosure, lawful recording consent, retention limits, access controls, authenticated webhooks, and avoid enabling HTTP mode on a reachable interface unless it is protected by network controls and request limits.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
mcp-server/server.py:251
Finding

Unauthenticated HTTP Service Permits Resource-Exhaustion Denial of Service

Content
View full analysis
Remediation
View remediation
MAX_BODY_SIZE: self.send_error(413, "Request body too large") return ``` 3. Configure connection and read timeouts so incomplete or slow requests cannot hold the service indefinitely. 4. Place remotely exposed deployments behind a production reverse proxy that enforces body-size limits, connection limits, rate limiting, TLS, and request timeouts. 5. Require authentication and authorization when the MCP endpoint must be remotely accessible. 6. Restrict requests to the intended endpoint and expected content type, rejecting unsupported paths and non-JSON requests. 7. Use a bounded concurrent server architecture so one slow client cannot prevent all legitimate requests from being processed. Concurrency limits must also be applied to avoid replacing single-request blocking with unbounded thread exhaustion. 8. Limit network access through host firewall rules, private-network binding, or an authenticated gateway. ]]>
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding

The skill declares a runtime dependency on gh and references local knowledge files, but it does not explicitly declare any tool scope or permissions. That creates ambiguity about what capabilities the skill may use at runtime and can lead to unintended file access beyond what reviewers or operators expect.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The trigger conditions are broad enough to activate on many generic voice AI or phone-automation questions, not just narrowly scoped WhoBot requests. Over-broad activation can cause unintended skill routing, increase prompt-surface exposure to untrusted content, and override a more relevant or safer skill for unrelated user queries.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

This code file includes user-facing natural-language documentation such as the module docstring and usage notes in Chinese, but does not indicate that users may choose another language. Under the language/locale policy, forcing a specific language without opt-in is a policy concern unless the locale restriction is explicit and justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The document explicitly markets the system as indistinguishable from a human caller, without any disclosure or opt-in framing. In a voice-agent skill, this is materially risky because it encourages deceptive interaction patterns that can mislead call recipients, undermine consent, and increase legal and reputational exposure when used in outbound or service contexts.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The knowledge file promotes call recording, transcription, CRM/webhook integration, and real-time business actions, but does not pair those capabilities with any disclosure, consent, retention, or access-control guidance. In a skill meant to answer product questions, this omission can normalize unsafe handling of sensitive customer data and lead downstream users to deploy privacy-impacting workflows without adequate safeguards.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The instruction '中文为主,英文为辅' imposes a language preference by default and does not indicate that the user can choose another language. This can conflict with language/locale policy expectations when no user opt-in or choice is offered.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The natural-language comment at L03 is written only in Chinese and provides required setup instructions without any opt-in, alternative language, or stated region-specific reason. This can violate a language/locale policy when skills are expected to avoid forcing a specific language on users by default.

Content

No source excerpt is available for this finding.