Back to skill

Security audit

Systematic Debugging

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward debugging checklist with broad trigger wording but no hidden code, persistence, credential access, or unsafe automatic behavior.

Installers should be aware that this skill may trigger on common debugging words and steer the assistant into a structured troubleshooting process. That is generally appropriate for software support, but publishers could narrow the trigger rules to reduce accidental activation.

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 (5)

Anti-Refusal Statement

High
Category
Anti-Refusal
Confidence
85% confidence
Finding

Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Content

Scanner excerpt · README.md (reported line 21)May include surrounding context.

md
- Reproduce the issue reliably.

### Phase 2: Hypothesize
- List all possible causes (brainstorm, don't judge yet)
- Rank by likelihood
- Identify what evidence would prove/disprove each

Vague Triggers

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The front-matter trigger list is broad enough to activate on many ordinary support or conversational requests such as 'error', 'bug', or 'fix this' without clear gating conditions. In an agentic system, overly broad activation can cause unintended tool/skill invocation, steering the assistant into a debugging workflow when the user did not explicitly request it, which can lead to inappropriate guidance or expanded access to project context.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The listed triggers include generic phrases such as "debug," "bug," "crash," "not working," and "troubleshoot," which are common in ordinary developer conversation and could cause unintended invocation. The README also provides no constraints, context limits, or negative examples to clarify when the skill should or should not activate.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The explicit trigger section lists short, ambiguous phrases like 'broken', 'debug', and 'fails' but does not define contextual limits, exclusions, or a confirmation step. This increases the chance of false-positive routing and unintended skill execution across unrelated domains, which is especially risky in assistants that automatically select skills based on phrase matches.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
78% confidence
Finding

The trigger list mixes English and Chinese activation phrases, implying language-specific behavior is embedded in the skill definition. While not inherently unsafe, the file does not explain locale selection, opt-in, or whether multilingual triggering is intentional policy, which can create ambiguity around language handling expectations.

Content

No source excerpt is available for this finding.