The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides instructions to ingest and process data from potentially untrusted internal sources such as Slack messages, corporate emails, and Google Drive documents. This creates a surface for indirect prompt injection attacks where malicious instructions embedded in a message or document could be executed by the agent.
Ingestion points: Access to Slack channels, email content, Google Drive files, and calendar events as detailed in examples/3p-updates.md, examples/company-newsletter.md, and examples/faq-answers.md.
Boundary markers: No specific boundary markers or instructions to ignore embedded directives are provided in the skill guidelines.
Capability inventory: The skill utilizes tools for reading communication platforms and document storage services.
Sanitization: No instructions for sanitizing or filtering the content retrieved from these external sources are included.
[DATA_EXFILTRATION]: The skill instructions direct the agent to aggregate sensitive corporate information from Slack, Email, and Google Drive into centralized reports. While intended for internal reporting, this consolidation of high-value internal data represents an information exposure risk if the generated content is shared outside the intended audience.

internal-comms