bearer-token-validator

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
  • Indirect Prompt Injection (HIGH): The skill is designed to process untrusted external data in a highly sensitive context.
  • Ingestion points: Processes user-provided authentication tokens, API configuration files, and source code for 'validation'.
  • Boundary markers: None present; the skill lacks delimiters or instructions to ignore embedded commands within processed data.
  • Capability inventory: Possesses Read, Write, Edit, and Bash(curl:*) permissions.
  • Sanitization: No evidence of input validation or output sanitization.
  • Risk: An attacker could provide a malicious payload disguised as a token that instructs the agent to read local secrets (e.g., ~/.aws/credentials) and exfiltrate them using the provided curl tool.
  • Data Exposure & Exfiltration (MEDIUM): The skill is explicitly granted Bash(curl:*) access without domain restriction. Given its purpose involves handling 'Bearer Tokens', this authorization is overly permissive and facilitates the silent transmission of credentials to external servers under the guise of validation checks.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 06:28 AM