Back to skill

Security audit

paper-summary-assistant

Security checks for vulnerabilities and agentic risk

Overview

This is a simple academic-paper summarization skill with no code execution, persistence, credential access, or hidden behavior.

This skill appears safe to install for structured academic paper summaries. Users should treat its outputs as study assistance and verify important paper details, especially because the template may be applied broadly to nearby academic tasks.

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
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Anti-Refusal Statement

High
Category
Anti-Refusal
Confidence
96% confidence
Finding

The phrase 'Always answer using' is an anti-refusal pattern because it attempts to force output behavior regardless of higher-priority safety, policy, or contextual constraints. In a benign educational skill this is likely intended as formatting guidance, but it can still reduce the model's willingness to refuse unsafe requests or to deviate when the required format is inappropriate, making instruction hierarchy conflicts more dangerous.

Content

Scanner excerpt · SKILL.md (reported line 83)May include surrounding context.

md
## Output format

Always answer using:

# Paper Summary

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The metadata description says to use the skill when users upload papers or ask for academic paper summaries, which is broad enough to activate in loosely related contexts without stronger boundaries. While not directly unsafe, overbroad activation can cause the skill to inappropriately take over conversations and apply its instructions where they do not fit, increasing the chance of instruction conflicts or misuse.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The 'When to use this skill' section provides only positive triggers and lacks scope limits or negative examples, so the skill may activate for adjacent tasks like presentations, article explanations, or generic document analysis. This can cause unintended invocation and make the assistant follow the skill in contexts where a different policy, tool, or workflow should apply.

Content

No source excerpt is available for this finding.