The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (HIGH): The documentation instructs users to install @anthropic-ai/gemini-cli. This is a critical security red flag as Gemini is a Google product, not an Anthropic one. This discrepancy is a strong indicator of a dependency confusion or typosquatting attack targeting users who expect official tools.
[EXTERNAL_DOWNLOADS] (MEDIUM): The skill is installed from an untrusted personal GitHub repository (rajivmehtaflex/gemini-search-skill) rather than a verified organization.
[COMMAND_EXECUTION] (LOW): The skill executes a shell script (scripts/search.sh) that triggers the gemini CLI. While the script uses double-quotes to prevent basic shell breakout, it remains a surface for executing system-level commands based on model output.
[PROMPT_INJECTION] (LOW): The skill is vulnerable to indirect prompt injection (Category 8).
Ingestion points: User-provided search queries are passed directly into the execution command in SKILL.md and search.sh.
Boundary markers: Absent. There are no delimiters or instructions to help the agent distinguish between a search query and embedded malicious instructions.
Capability inventory: The skill possesses the capability to execute shell commands and access the network via the gemini CLI.
Sanitization: None detected. The raw query string is interpolated directly into the command line.

gemini-search