xray
Warn
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses user-supplied arguments to construct file paths for shell commands, such as
ls "$MEMORY_DIR"/*<name>*.mdandmv. If a user provides input containing shell metacharacters (e.g., semicolons, backticks, or pipes), it could lead to arbitrary command execution on the host system. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it reads and displays the contents of memory files from
$HOME/.claude/projects/which are populated with data from various external sources during agent interactions. 1. Ingestion points: Markdown files in the project memory directory. 2. Boundary markers: No explicit delimiters or instructions are present to prevent the agent from executing instructions found within the data. 3. Capability inventory: The skill utilizes shell commands for file listing, reading, and moving. 4. Sanitization: Memory file content is not sanitized or escaped before being processed in the agent context.
Audit Metadata