The Agent Skills Directory

Command Execution: The skill executes local utilities (checker, redactor) using the uv run command. These tools are located within the framework's own directory structure and are used to verify the privacy configuration and redaction capabilities.
Indirect Prompt Injection Surface: The skill processes data (such as mailing list content) through a redaction tool. This represents a potential surface where malicious instructions embedded in the processed data could attempt to influence the agent's behavior.
Ingestion points: The pii-redact tool reads data from standard input in the verification step (SKILL.md, Step 4).
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are specified for the input data during the verification process.
Capability inventory: The skill uses uv run to execute local scripts and tools (SKILL.md, Step 4).
Sanitization: The process relies on the pii-redact tool to sanitize data by removing personally identifiable information before it reaches an LLM.

privacy-llm