elephant-agent-personal-ai
Fail
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: CRITICALREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill provides an installation command that pipes a remote shell script directly into bash (
curl -fsSL https://elephant.agentic-in.ai/install.sh | bash). This practice is dangerous because it allows unverified code execution with the user's current privileges. - [COMMAND_EXECUTION]: The provided Python example for a 'GitSkill' uses the
subprocessmodule to execute system commands (subprocess.run). This demonstrates a capability to run arbitrary shell commands on the system. - [EXTERNAL_DOWNLOADS]: The skill references downloads from external sources, including a
.dmgfile from GitHub (github.com/agentic-in/elephant-agent) and a shell script fromelephant.agentic-in.ai. - [CREDENTIALS_UNSAFE]: The documentation includes example environment variables with API key prefixes (
sk-...,sk-ant-...). These markers are identified as unsafe credential patterns. - [PROMPT_INJECTION]: The skill's 'FileSystemSkill' and 'BrowserSkill' ingest untrusted external data (local files and web content), creating a surface for indirect prompt injection (Category 8).
- Ingestion points: Local file access through
FileSystemSkilland web scraping viaBrowserSkill. - Boundary markers: No specific boundary markers or 'ignore' instructions are provided in the examples.
- Capability inventory: The skill shows examples of executing shell commands (
subprocess.run), making it vulnerable to privilege escalation if injected instructions are followed. - Sanitization: No sanitization or validation of the input data is demonstrated in the provided code snippets.
Recommendations
- HIGH: Downloads and executes remote code from: https://elephant.agentic-in.ai/install.sh - DO NOT USE without thorough review
- AI detected serious security threats
Audit Metadata