hunt-idor
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides numerous example commands for CLI tools such as curl, ffuf, and grep. These are standard tools for security professionals and are intended to be used on authorized targets to validate object-level authorization.
- [DYNAMIC_EXECUTION]: A Python code snippet is included to automate the generation of numeric ID wordlists. This script performs simple local file operations by writing a range of identifiers to a file named 'ids.txt'.
- [INDIRECT_PROMPT_INJECTION]: The skill's core functionality involves analyzing responses from external web targets, which creates a surface for indirect prompt injection if those targets provide malicious content.
- Ingestion points: API response bodies, HTTP headers, and static JavaScript assets from target systems are processed by the agent.
- Boundary markers: The skill defines a 'Mandatory confirmation gate' that prevents the agent from performing active probes or data extraction until the user confirms the target, scope, and authorization.
- Capability inventory: The agent uses shell command execution (curl, ffuf) and local file writing as part of its security testing workflow.
- Sanitization: There are no explicit instructions for the agent to sanitize or escape data retrieved from target systems before processing it.
Audit Metadata