The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands for project validation.
Evidence: In rules/rust-validation.md, the skill directs the user/agent to run yarn check:all, cargo check, or cargo clippy after making changes.
Context: While these are standard tools, yarn scripts are defined in a project's package.json, which could be configured to execute malicious commands if the project files are untrusted.
[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for processing various project files, creating an attack surface for indirect prompt injection.
Ingestion points: The skill processes Rust source files (.rs), SQL files (.sql), and prompt files (.prompt) within the src/, sql/, and prompts/ directories.
Boundary markers: There are no instructions provided to delimit or treat the content of these files as untrusted data.
Capability inventory: The skill operates in an environment where the agent is expected to read/write files and execute local commands.
Sanitization: No sanitization or validation logic is present for the content extracted from project files before it is processed by the agent.

busirocket-rust