secret-management-attack
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted content from multiple sources, which could contain malicious instructions designed to manipulate the agent's logic.\n
- Ingestion points: Data enters the agent's context through repository cloning (SKILL.md), web crawling via the
caridditool (payloads.md), and Kubernetes secret enumeration (payloads.md).\n - Boundary markers: The instructions do not specify the use of clear delimiters or boundary markers when processing external data, which is a key control for preventing prompt injection.\n
- Capability inventory: The skill has significant capabilities including shell command execution (
Bash) and file system manipulation.\n - Sanitization: There are no explicit sanitization or filtering steps defined for content retrieved from scanned repositories or web pages before it is analyzed by the agent.\n- [COMMAND_EXECUTION]: The skill provides a comprehensive library of shell commands for performing secret discovery, verifying tokens, and exploiting platform misconfigurations.\n- [EXTERNAL_DOWNLOADS]: The skill fetches several auditing tools (gitleaks, trufflehog, bearer, and DeepAudit) from their respective GitHub repositories or release pages.
Audit Metadata