The Agent Skills Directory

COMMAND_EXECUTION (MEDIUM): The skill definition utilizes a shell command (grep) to extract dependency information from the Cargo.toml file. Executing shell commands directly from skill instructions poses a risk if the execution environment is not strictly sandboxed.
PROMPT_INJECTION (HIGH): Indirect Prompt Injection vulnerability (Category 8). The skill ingests content from Cargo.toml, an external file that could be controlled by an attacker (e.g., via a malicious pull request or a compromised project). This content is provided to the agent as context without protective measures.
Ingestion points: Cargo.toml (read via the !grep shell command).
Boundary markers: Absent; the output of the command is injected directly into the agent's prompt context without delimiters or instructions to ignore embedded commands.
Capability inventory: The skill's stated purpose involves "integrating crates", "FFI integration", and "dependency management", which strongly implies the agent has permissions to modify project files and execute build tools (e.g., cargo).
Sanitization: Absent; no escaping or validation is performed on the data read from the file.

m11-ecosystem