The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill identifies and ingests content from the local workspace which may contain untrusted data or malicious instructions.
Ingestion points: Uses rg (ripgrep) in SKILL.md to search and read content from Cargo.toml, Cargo.lock, *.rs, and *.sql files within the repository.
Boundary markers: The skill does not provide explicit delimiters or instructions to the agent to treat the ingested file content as untrusted data separate from its primary instructions.
Capability inventory: The skill invokes cargo test, cargo sqlx, and cargo deny. While these are standard tools, executing cargo test on a workspace with maliciously modified test files could lead to local code execution.
Sanitization: There is no evidence of sanitization or filtering of the file content retrieved by search commands before it is processed by the agent.

rust-database