kernel-audit-security
Warn
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DYNAMIC_EXECUTION]: The methodology in 'audit-methodology/SKILL.md' and 'audit-methodology/MEMORY-SAFETY-AND-BINARY.md' explicitly directs the agent to build and run the target codebase or extract suspect code into a harness for verification. This practice is inherently risky when applied to untrusted or potentially malicious codebases.
- [INDIRECT_PROMPT_INJECTION]: As an audit tool, the skill's primary function is to ingest and reason about large volumes of untrusted data. Malicious codebases could contain indirect prompt injections designed to subvert the auditing agent, especially since it has the capability to write files and execute shell commands.
- [COMMAND_EXECUTION]: The skill relies on shell commands for reconnaissance and includes internal scripts like 'validate-findings.cjs' that the agent is instructed to run.
- [EXTERNAL_DOWNLOADS]: The skill contains update instructions that fetch security standards and framework data from public GitHub repositories. These are well-known resources used for data enrichment.
Audit Metadata