The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is instructed to process and summarize untrusted data from various external communication and documentation tools.
Ingestion points: The guidelines in examples/3p-updates.md, examples/company-newsletter.md, and examples/faq-answers.md direct the agent to gather context from Slack messages, Google Drive documents, company emails, and calendar events.
Boundary markers: The skill lacks explicit instructions or delimiters that would prevent the agent from following instructions embedded within the ingested messages or documents.
Capability inventory: The agent uses the gathered data to create internal communications that are distributed to other employees, executives, and leadership, potentially spreading malicious content.
Sanitization: There is no defined process for validating or sanitizing external content to ensure it does not contain adversarial instructions.

internal-comms