The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a shell command that directly interpolates the $ARGUMENTS variable: node ${CLAUDE_PLUGIN_ROOT}/src/export.js $ARGUMENTS. This lacks shell quoting (e.g., "$ARGUMENTS") or input validation, allowing an attacker or malicious data to execute arbitrary shell commands by including metacharacters like ;, &, |, or backticks in the input.\n- [PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection by passing untrusted input directly into a shell environment.\n
Ingestion points: The $ARGUMENTS variable in SKILL.md.\n
Boundary markers: None. The input is used directly without delimiters or instructions to ignore embedded commands.\n
Capability inventory: The skill uses the Bash tool to execute shell commands via node.\n
Sanitization: None. The variable is neither escaped nor quoted.

cco-export