The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes standard development tools including rust-analyzer, cargo, and jq. The provided script scripts/list-crate-features.sh interacts with the local workspace metadata to help the agent or user select appropriate feature flags and packages for analysis.
[DATA_EXFILTRATION]: No network exfiltration patterns or unauthorized access to sensitive file paths (such as credentials or SSH keys) were detected. The operations are confined to the local workspace and standard Rust toolchain environment.
[EXTERNAL_DOWNLOADS]: The skill does not perform any external downloads or execute remote scripts. It relies on locally installed development tools.
[PROMPT_INJECTION]: The instructions are purely technical and task-oriented, focusing on code analysis workflow without any attempts to override safety filters or bypass system instructions.
[INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes workspace metadata and source code.
Ingestion points: Workspace Cargo.toml files and Rust source files are processed via cargo metadata and rust-analyzer in SKILL.md and scripts/list-crate-features.sh.
Boundary markers: None explicitly present in the bash script output format, but the data is structured as JSON via jq before processing.
Capability inventory: The skill can execute rust-analyzer and cargo commands in SKILL.md and references/rust-analyzer-cli.md.
Sanitization: The helper script uses jq's --arg and --argjson flags to safely handle package names and features, preventing command injection during the generation of suggested commands.

risingwave-rust-analyzer