ClawHub
Back to skill

Security audit

Authy

Security checks for vulnerabilities and agentic risk

Overview

This skill is a documented wrapper for injecting scoped secrets into subprocesses, with no hidden code or persistence, but users should only run it with trusted commands.

Install only if you intend agents to run trusted local commands with scoped secrets. Review each authy run command, use the narrowest scope available, and avoid injecting secrets into scripts or network clients unless you trust the script and the destination.

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

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill description is broad enough to match many generic requests involving API keys, credentials, or tokens, which increases the chance the agent will invoke this skill in situations where secret injection is unnecessary or inappropriate. Because the skill is specifically designed to place secrets into subprocess environments, over-triggering expands the attack surface for accidental disclosure or misuse.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
90% confidence
Finding

The included curl example demonstrates direct external transmission of a secret-bearing Authorization header to a remote endpoint. In the context of a secret-injection skill, this is more dangerous because the skill's core function is to make secrets available to subprocesses, so any network-capable command can immediately exfiltrate them.

Content

Scanner excerpt · SKILL.md (reported line 55)May include surrounding context.

bash
cat > task.sh << 'EOF'
#!/bin/bash
curl -H "Authorization: Bearer $API_KEY" https://api.example.com/data
EOF
chmod +x task.sh
authy run --scope my-scope --uppercase --replace-dash '_' -- ./task.sh

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The example script sends an Authorization bearer token to a remote HTTPS endpoint without any warning that the secret will leave the local process boundary and be disclosed to an external service. Even if this is a legitimate use case, presenting it as a default example can normalize unsafe behavior and cause users or agents to transmit injected secrets to unintended destinations.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.