The Agent Skills Directory

[PROMPT_INJECTION] (LOW): The skill is vulnerable to Indirect Prompt Injection (Category 8). An attacker could embed malicious search queries in data processed by the agent (e.g., a document or email) to trick it into searching for sensitive system information.
Ingestion points: User-supplied search queries passed as command-line arguments to scripts/search.js.
Boundary markers: None. The script does not use delimiters or instructions to prevent the execution of embedded search commands within the user input.
Capability inventory: The skill uses spawnSync to call es.exe, providing full read-only visibility into the filesystem, including filenames, paths, and file content (via the content: query).
Sanitization: No sanitization or validation is performed on the query arguments before they are passed to the external binary.
[COMMAND_EXECUTION] (LOW): The script scripts/search.js executes an external binary (es.exe). It correctly uses shell: false in spawnSync, which prevents standard shell injection attacks (like command chaining with ; or &). However, the execution of external binaries is a high-privilege operation that is core to this skill's functionality.

everything-search