Security audit

Feishu CRM Lite

Security checks for vulnerabilities and agentic risk

Overview

The skill is a CRM helper, but it under-discloses that customer data is stored locally in plaintext and its documentation overstates Feishu integration that is not implemented.

Install only if you are comfortable with customer and follow-up data being kept as local plaintext files under data/feishu-crm, not in Feishu Bitable as documented. Run it from a private workspace, avoid adding regulated or highly sensitive customer details, and require explicit confirmation before deleting records or doing bulk import/export until the storage model and Feishu integration are clarified.

A.I.G

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: index.js:53
Finding: CRM Personally Identifiable Information Stored in Plaintext Files with Implicit Permissions
Content: View full analysis
c.status === filters.status); } if (filters.tag) { customers = customers.filter(c => c.tags.includes(filters.tag)); } if (filters.source) { customers = customers.filter(c => c.source === filters.source); } return customers; } saveCustomers(customers) { fs.writeFileSync(this.customersFile, JSON.stringify(customers, null, 2)); } ``` ```js getFollowUps(customerId = null) { const data = fs.readFileSync(this.followUpsFile, 'utf8'); let followUps = JSON.parse(data); if (customerId) { followUps = followUps.filter(f => f.customerId === customerId ...[truncated 3697 chars]
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (11)

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The README states the skill can be used immediately with no extra configuration, while the development checklist later shows core CRM, follow-up, reminder, funnel, and analytics features are not yet implemented. This mismatch can mislead users into trusting and invoking actions that do not exist or may behave unpredictably, increasing the chance of operational mistakes or unsafe assumptions about data handling.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The skill advertises potentially state-changing operations such as creating, updating, deleting, importing/exporting customer records, and setting reminders without any warning about modifying user data or triggering external effects. In an agent context, lack of disclosure and confirmation guidance can cause users to invoke destructive or privacy-sensitive actions without understanding the consequences.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The trigger list includes very broad everyday terms such as '提醒', '销售', '客户管理', and 'customer', which increases the chance the skill activates in unrelated conversations. In a skill that claims the ability to create, update, delete, and store customer records, accidental invocation could lead to unintended data access or modification.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The skill description describes customer record creation, update, deletion, import/export, and storage in Feishu Bitable without warning users that business contact data will be persisted and modified. Because CRM data often contains personal and commercial information, missing consent, retention, and modification warnings increases privacy and compliance risk.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The skill says it archives communication content and uses existing Feishu API permissions, but it does not disclose the privacy implications of storing conversation details or the scope of access granted through those permissions. Archiving follow-up notes can capture sensitive personal, commercial, or regulated information, making silent collection and retention risky.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The skill advertises Feishu CRM capabilities such as Bitable-backed storage, customer management, follow-up tracking, reminders, and analytics as available features, while the development status explicitly says those core integrations and functions are not yet implemented. This can mislead users into relying on nonexistent functionality for customer data handling or business workflows, creating integrity and operational risk.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: This code initializes and writes persistent files for customers, follow-ups, and reminders, which contain potentially sensitive business contact information. Although the file has code comments, there is no user-facing disclosure, confirmation, or CLI warning that customer data will be stored on disk under the current working directory.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 88% confidence
Finding: The deleteCustomer method irreversibly removes customer records from persisted storage, but there is no confirmation prompt, warning message, or explicit disclosure of this destructive behavior. Code comments alone do not warn end users, and this operation could result in accidental data loss if exposed through a skill action.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 87% confidence
Finding: SQP-3 applies to natural-language policy issues in all file types. The file presents all instructions and examples only in Chinese and does not indicate that the user can choose another language or that the skill is intentionally limited to a Chinese-speaking or region-specific audience.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 80% confidence
Finding: All natural-language instructions, headings, and examples are presented in Chinese, and the file does not indicate that the skill is intentionally limited to Chinese-speaking users or a specific locale. Under the stated policy, forcing a specific language without user opt-in or documented justification is a natural-language policy concern.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The natural-language description and CLI output strings are Chinese-only, which can impose a language/locale choice on users without opt-in. The policy calls for flagging skills that force a specific language unless they clearly offer a choice or justify the locale restriction.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.