The Agent Skills Directory

[SAFE]: The skill establishes a security-conscious workflow for Rust development, specifically requiring documentation of invariants for unsafe blocks and minimizing visibility of FFI boundaries.- [COMMAND_EXECUTION]: The workflow requires execution of repository-standard development tools including cargo, rustfmt, and clippy. These are expected operations for a Rust implementation skill and occur within the local repository context.- [EXTERNAL_DOWNLOADS]: The skill utilizes official package registries (crates.io) via the standard cargo toolchain. External references to repositories like oven-sh/bun are for documentation and parity analysis purposes and originate from well-known sources.- [PROMPT_INJECTION]: The skill identifies ingestion points for repository-provided guidance (e.g., AGENTS.md, CLAUDE.md, and architecture docs). While this represents a surface for indirect prompt injection, it is a functional requirement for the skill to align with project-specific standards, and the instructions prioritize existing interfaces and safety contracts.

implement-rust