erf-dump

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a PowerShell script using 'powershell.exe' with arguments dynamically constructed from user input and configuration files.
  • [CREDENTIALS_UNSAFE]: The skill instructions direct the agent to handle database passwords and pass them as plaintext command-line arguments, which can be captured in process logs and shell history.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection. 1. Ingestion points: Data is read from '.v8-project.json' in the skill body. 2. Boundary markers: No delimiters are used to isolate untrusted data. 3. Capability inventory: The skill performs shell execution via PowerShell scripts. 4. Sanitization: No sanitization or validation of input from the configuration file is performed.
  • [EXTERNAL_DOWNLOADS]: The skill relies on a script from a sibling directory ('../epf-dump/scripts/epf-dump.ps1'), creating a dependency on code that is not contained or verified within this specific skill folder.
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 11:10 AM
Security Audit — agent-trust-hub — erf-dump