The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The file SKILL.md contains the command cat .cargo/config.toml 2>/dev/null || echo "No .cargo/config.toml found" using the !command syntax. This mechanism automatically executes the command at skill load time to provide the AI agent with the project's target configuration and build environment details. This is a standard and benign use case for developer-oriented tools.
[INDIRECT_PROMPT_INJECTION]: The skill ingests external project data into the agent's context, creating a potential surface for indirect instructions if the project file is maliciously crafted.
Ingestion points: The contents of .cargo/config.toml are injected into the Project Context section of SKILL.md via a dynamic shell command.
Boundary markers: There are no explicit delimiters or instructions telling the model to ignore potential commands within the injected file content.
Capability inventory: The skill itself does not define any additional executable scripts, high-privilege tool access, or network capabilities.
Sanitization: No sanitization or validation is performed on the content of the configuration file before it is presented to the agent.

domain-embedded