Security audit

Web Search Rules

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed, high-capability workflow for saving web search results into local or cloud knowledge bases with confirmation controls.

Install this only if you want an agent to manage search-result capture into a knowledge base. Prefer local staging first, review whitelist/blacklist rules before reuse, and enable cloud upload, browser automation, deletion, or migration only after confirming the target platform and item list.

A.I.G

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

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
Rogue AgentSelf-Modification, Session Persistence
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (23)

Ae1

High

Category: analysis-evasion
Confidence: 100% confidence
Finding: Referenced artifact was not completely inspected
Content: Scanner excerpt · SKILL.md (reported line 79)May include surrounding context.
md
Read `references/platform-adapters.md` before using any adapter-specific operation.

Ae1

High

Category: analysis-evasion
Confidence: 100% confidence
Finding: Referenced artifact was not completely inspected
Content: Scanner excerpt · SKILL.md (reported line 192)May include surrounding context.
md
Read `references/platform-adapters.md` before using any adapter-specific operation.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · SECURITY.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · SKILL.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · references/examples.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · references/feishu-dingtalk-operations.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · references/ima-operations.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · references/migration-and-testing.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · references/notebooklm-operations.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · references/obsidian-operations.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · references/platform-adapters.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · references/platform-comparison.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · references/rule-engine.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Hidden Instructions

High

Category: Prompt Injection
Confidence: 60% confidence
Finding: Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
Content: Scanner excerpt · references/tencent-docs-operations.md (reported line 1)May include surrounding context.
md
# Tencent Docs Operations Tencent Docs is a cloud collaborative document adapter. Treat rule files and staged content as cloud data.

Session Persistence

Medium

Category: Rogue Agent
Confidence: 60% confidence
Finding: Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
Content: Scanner excerpt · SKILL.md (reported line 15)May include surrounding context.
md
## Security Notice This skill can guide an agent to read and write knowledge-base content, automate browsers, upload selected content to cloud platforms, and perform deletion or migration operations after confirmation. Read `SECURITY.md` before using it. Default safety posture:

Session Persistence

Medium

Category: Rogue Agent
Confidence: 60% confidence
Finding: Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
Content: Scanner excerpt · references/examples.md (reported line 15)May include surrounding context.
md
4. Apply rules. 5. Stage whitelisted and pending results. 6. Ask the user to choose whitelist, blacklist, save, or ignore. 7. Write confirmed changes and append audit logs. Report template:

Autonomous Decision Making

Medium

Category: Excessive Agency
Confidence: 85% confidence
Finding: Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
Content: Scanner excerpt · references/examples.md (reported line 25)May include surrounding context.
md
Platform: obsidian Total results: 18 Deduplicated: 14 Auto-approved: 3 Blacklisted: 2 Pending confirmation: 9 Saved: 5

Autonomous Decision Making

Medium

Category: Excessive Agency
Confidence: 85% confidence
Finding: Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
Content: Scanner excerpt · references/rule-engine.md (reported line 60)May include surrounding context.
md
Platform: obsidian Total results: 18 Deduplicated: 14 Auto-approved: 3 Blacklisted: 2 Pending confirmation: 9 Saved: 5

Persistent Context Injection

Medium

Category: Memory Poisoning
Confidence: 80% confidence
Finding: Skill injects content designed to persist in agent memory or context across interactions. Persistent injection can alter agent behavior long after the initial interaction.
Content: Scanner excerpt · references/examples.md (reported line 43)May include surrounding context.
Domain: example.com Reason: 6 previously saved items from this domain Proposed action: whitelist domain Options: apply for this run only, create persistent rule, keep reviewing one by one

text
## Cleanup dry-run

Autonomous Decision Making

Medium

Category: Excessive Agency
Confidence: 85% confidence
Finding: Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
Content: Scanner excerpt · SKILL.md (reported line 33)May include surrounding context.
md
Allowed actions: - `whitelist`: auto-stage and mark auto-approved. - `blacklist`: skip by default and report as filtered. - `uncategorized`: stage for user review. - `needs_review`: stage only summary and ask before fetching full content.

Autonomous Decision Making

Medium

Category: Excessive Agency
Confidence: 85% confidence
Finding: Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
Content: Scanner excerpt · references/rule-engine.md (reported line 32)May include surrounding context.
md
Allowed actions: - `whitelist`: auto-stage and mark auto-approved. - `blacklist`: skip by default and report as filtered. - `uncategorized`: stage for user review. - `needs_review`: stage only summary and ask before fetching full content.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 77% confidence
Finding: The file identifies the skill as web-search-rules-en, which indicates an English-specific locale, but the document does not state that language selection is optional or explain why English is required. Under the policy, locale constraints should be opt-in or clearly justified.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 79% confidence
Finding: The statement 'Obsidian is the preferred local adapter' imposes a tool preference in natural language without indicating user opt-in or alternatives. Under the policy-violation category, prescriptive language that forces a specific choice can be problematic when no explicit choice or justification is provided.
Content: No source excerpt is available for this finding.