Back to skill

Security audit

Systematic Debugging

Security checks for vulnerabilities and agentic risk

Overview

This debugging skill is mostly guidance and helper examples; it has some sensitive diagnostic examples users should run only when appropriate.

Install only if you want a strict root-cause-first debugging workflow. Treat included helper scripts as normal command runners: they do not make arbitrary commands safe. If the agent proposes keychain, signing, environment, or secret-related diagnostics, approve only the minimum needed command and redact output before sharing logs.

Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The docstring explicitly describes the tool as 'Safe' and says it runs only the provided command with 'no extra mutations,' which can mislead users into believing destructive commands are somehow constrained or harmless. In reality, the script will execute any command passed to it, so the safety claim can encourage risky use and accidental destructive execution.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This prompt uses urgency, financial pressure, and explicit instructions to 'choose and act' in a way that pushes an agent or user toward operational decisions without adequate validation. In a skill context, that is dangerous because it conditions bypassing normal diagnostic safeguards and can lead to unsafe production changes based on incomplete evidence.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The scenario explicitly pressures rapid deployment of a production fix while discouraging investigation, emphasizing revenue loss and reputational consequences to bias the decision. This is risky because it normalizes making live system changes under pressure without sufficient root-cause analysis, testing, or rollback planning, increasing the chance of outages, regressions, or security-impacting misconfigurations.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
87% confidence
Finding

The skill includes diagnostic commands that enumerate keychain state and signing identities. Even though intended for troubleshooting, this can expose sensitive environmental information about available credentials and security configuration, which may aid an attacker in credential discovery or post-exploitation reconnaissance.

Content

Scanner excerpt · SKILL.md (reported line 184)May include surrounding context.

md
env | grep IDENTITY || echo "IDENTITY not in environment"

   # Layer 3: Signing script
   echo "=== Keychain state: ==="
   security list-keychains
   security find-identity -v

Credential Access

High
Category
Privilege Escalation
Confidence
89% confidence
Finding

The explicit use of security find-identity -v is a form of credential and identity enumeration. In a generic skill that may run in varied environments, this increases risk because it teaches the agent to probe local signing assets and could disclose certificate identities or metadata beyond what is necessary for debugging.

Content

Scanner excerpt · SKILL.md (reported line 185)May include surrounding context.

md
# Layer 3: Signing script
   echo "=== Keychain state: ==="
   security list-keychains
   security find-identity -v

   # Layer 4: Actual signing