The Agent Skills Directory

[SAFE]: The skill contains no obfuscated code, hidden characters, or deceptive metadata. All instructions are clear and aligned with the stated purpose of auditing Rust code.
[SAFE]: All external references in the corpus point to highly reputable and well-known projects within the Rust ecosystem, such as Tokio, Ripgrep, and Rusqlite. These are used for architectural reference rather than remote script execution.
[COMMAND_EXECUTION]: The skill instructs the agent to utilize standard Rust development tools (e.g., cargo check, cargo clippy, cargo audit, cargo test) to validate audit hypotheses. While running tests on untrusted code can theoretically execute arbitrary code, these tools are standard for the primary purpose of code auditing and are invoked within a structured, read-only workflow described in the skill.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted repository data, which is an inherent surface for indirect prompt injection attacks.
Ingestion points: The agent is instructed to read repository manifests, entry points, source code, documentation, and specific guidance files like AGENTS.md and CLAUDE.md (SKILL.md, references/testing-and-quality.md).
Boundary markers: The skill establishes a priority hierarchy that incorporates repository-specific guidance but does not include explicit warnings to ignore instructions found within the audited files.
Capability inventory: The skill has the capability to read files and execute toolchain-level commands such as cargo check, cargo test, and cargo bench (references/testing-and-quality.md).
Sanitization: There are no explicit instructions for sanitizing or filtering external repository content before processing.

audit-rust