The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process arbitrary user-provided text to identify and redact PII, which introduces a surface for indirect prompt injection attacks where malicious instructions could be embedded in data processed by the agent.
Ingestion points: Text data is ingested via scan_and_redact in SKILL.md and PIIDetector.redact in references/examples.md.
Boundary markers: The skill provides a validate_request utility in references/security-best-practices.md that checks for common injection patterns like 'ignore previous instructions'.
Capability inventory: The skill has access to the Bash tool and performs network requests to the OpenRouter API via the openai library.
Sanitization: The skill performs PII redaction using regular expressions to replace sensitive values with placeholders, which acts as a form of content sanitization.
[SAFE]: The skill demonstrates security best practices by recommending the use of environment variables for API keys and implementing audit logging that only stores SHA-256 hashes of prompts rather than the raw content. The instructions to use .env files and environment variables for secrets are considered safe standard practices.

openrouter-data-privacy