The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified in the source files. The tool performs legitimate prompt optimization tasks such as whitespace stripping, comment removal, and context truncation to reduce token consumption.\n- [SAFE]: Credential management in main.rs and config/mod.rs is secure. API keys for various providers are handled via environment variables or local configuration files and are masked when displayed in terminal output.\n- [SAFE]: Network communication in api/client.rs and api/venice.rs is restricted to established and well-known service providers (Anthropic, OpenAI, Venice.ai) and local Ollama instances for preprocessing.\n- [SAFE]: Command execution in orchestrator/mod.rs is limited to the documented integration with the official Claude CLI for provider fallback purposes, which is a documented and intended feature.\n- [SAFE]: All library dependencies listed in Cargo.toml are standard, reputable Rust crates used for networking, serialization, and terminal UI components.

token-optimizer