The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for using standard development utilities including git log, cargo test, cargo clippy, and gh pr view. These tools are used for generating local test fixtures and performing quality gates. The operations are conducted within the scope of a local project directory.
[DATA_EXPOSURE]: The skill suggests capturing output from commands like git log and gh pr view into local fixture files (tests/fixtures/*.txt). This data is used solely for offline testing and snapshot verification within the project. No access to sensitive user credentials, private keys, or environment variables was detected.
[REMOTE_CODE_EXECUTION]: The workflow leverages common Rust libraries such as insta, regex, lazy_static, and anyhow. These are standard, well-maintained packages in the Rust ecosystem. The skill does not include patterns for downloading or executing arbitrary scripts from untrusted remote sources.
[INDIRECT_PROMPT_INJECTION]: The skill defines a process for ingesting external data from command outputs for filtering purposes.
Ingestion points: Data is read from the tests/fixtures/ directory via include_str! macro in the Rust test code.
Boundary markers: Snapshot testing using the insta crate provides a mechanism to lock and verify output formats.
Capability inventory: The skill execution environment includes the Bash tool and Rust functions capable of system command execution, which is standard for developer tooling.
Sanitization: Content is processed using regular expression filtering to extract specific log levels, which effectively limits the data passed into the application context.
[NO_CODE]: The skill serves as a set of instructional guidelines and code templates for developers to implement, rather than delivering pre-compiled binaries or hidden logic strings.

tdd-rust