The Agent Skills Directory

[SAFE]: The skill defines security best practices for Rust development, utilizing industry-standard tools like cargo-audit and cargo-deny to maintain project integrity.
[COMMAND_EXECUTION]: Employs standard Rust toolchain commands and shell utilities for dependency analysis and the removal of potentially malicious local files. All operations are strictly localized and aligned with the skill's security auditing purpose.
[EXTERNAL_DOWNLOADS]: Communicates with the official crates.io API to retrieve crate metadata and fetches the RustSec advisory database. These are recognized, well-known services within the technology domain.
[INDIRECT_PROMPT_INJECTION]: The skill manages a workflow involving the ingestion of third-party source code and manifests.
Ingestion points: Processes Cargo.lock files and external crate source directories as described in Section 5 of SKILL.md.
Boundary markers: Explicitly cautions against the use of automated IDE build scripts and requires manual vetting before code compilation or testing.
Capability inventory: Uses CLI tools such as cargo, grep, find, and curl for system-level analysis and maintenance.
Sanitization: Prescribes manual source code review using specific search patterns to identify suspicious behaviors like unauthorized network access or obfuscated payloads.

rust-security