code-quality-expert
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its core function of analyzing external, untrusted code provided by users.
- Ingestion points: User-supplied code blocks or file content ingested via the
Readtool or direct chat input for review according to instructions inSKILL.md. - Boundary markers: Absent. The skill instructions do not define clear delimiters or provide instructions to the model to ignore potential directives embedded within the code being analyzed.
- Capability inventory: The skill environment includes highly capable tools such as
Bash,Write,Edit,Read,Grep, andGlob. - Sanitization: No validation or sanitization of the input code is performed before the model interacts with it using available tools.
- [COMMAND_EXECUTION]: The skill explicitly instructs the agent to execute shell commands to manage its internal memory.
- Evidence: The
Memory ProtocolinSKILL.mdrequires the execution ofcat .claude/context/memory/learnings.mdat the start of the session to retrieve context.
Audit Metadata