Security audit

Content Collector Skill

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it says, but it can silently copy shared links and their contents into broadly visible Feishu storage without enough user control.

Install only if the workspace intentionally wants a globally enabled shared archive. Before enabling it, require explicit collection commands or admin-scoped opt-in groups, restrict readable and writable Feishu scopes, preserve source permissions, add URL and image-fetch safeguards, and define deletion/retention controls for documents, tables, indexes, uploaded images, and MEMORY.md logs.

A.I.G

Vulnerability Patterns

Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
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

Findings (2)

T05 · Unauthorized Access and Privilege Escalation

Error

Location: SKILL.md:29
Finding: Silent Cross-Group Collection and Globally Readable Archival Without Affirmative Consent
Content: View full analysis
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Error

Location: SKILL.md:247
Finding: Unrestricted Server-Side URL and Image Retrieval Can Expose Internal Resources
Content: View full analysis
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 (7)

Vague Triggers

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The skill defines 'silent collection' with broad triggers like any recognizable knowledge-type link in a group chat and no explicit per-message approval. This can cause unintended fetching, archiving, and redistribution of shared content based on ambiguous signals, especially when users may not realize their links will be copied into a persistent shared knowledge base.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The skill lacks a clear warning that it may automatically fetch external URLs, archive contents, and make resulting documents visible to all users. Without transparent notice, users cannot make informed decisions about whether to share links, creating privacy, confidentiality, and compliance risk for internal documents or licensed third-party content.
Content: No source excerpt is available for this finding.

Ssd 3

High

Category: Not specified by scanner
Confidence: 99% confidence
Finding: The skill instructs automatic collection of group-chat links, then stores source URLs, summaries, and content into shared Feishu documents and tables broadly visible to all users. In context, this is dangerous because chat-shared links may point to sensitive internal docs, private learning materials, or copyrighted content, and the workflow persists that data without clear per-message consent or access-boundary checks.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: Lines L112-L113 state that the skill must update only the configured knowledge base table and save documents under the designated knowledge base. However, lines L765-L782 introduce an additional capability to create records in an independent "图片资源库" Bitable, which directly contradicts the earlier documented restriction rather than merely extending it.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 89% confidence
Finding: The response templates are hardcoded under the zh_cn locale and the document repeatedly specifies Chinese-only output structures, but it does not offer any user opt-in or alternate language selection. This is a natural-language policy issue because it imposes a specific locale by default rather than allowing user choice.
Content: No source excerpt is available for this finding.

Ssd 3

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: Persistently logging every collection event and metadata into MEMORY.md creates an additional uncontrolled retention channel for user-shared links, titles, categories, and source information. Even if the main archive is intended, duplicating metadata into long-lived memory broadens exposure, complicates deletion, and may retain sensitive references beyond user expectations.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Low

Category: Not specified by scanner
Confidence: 89% confidence
Finding: The skill documentation repeatedly requires that all created documents be stored under the designated knowledge base using the wiki_node parameter. Yet the example at L738-L741 shows feishu_create_doc called with only title and markdown, contradicting the stated mandatory storage rule and potentially causing documents to be created outside the required location.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.