The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to proactively search and summarize data from potentially untrusted or attacker-controlled sources, which presents a surface for indirect prompt injection attacks.
Ingestion points: Slack messages, Google Drive documents, Emails, and Calendar event descriptions are identified as primary data sources in examples/3p-updates.md, examples/company-newsletter.md, and examples/faq-answers.md.
Boundary markers: The instructions lack explicit boundary markers or warnings to ignore malicious instructions embedded within the retrieved documents or messages.
Capability inventory: The skill uses these inputs to generate textual outputs like team updates, newsletters, and FAQs, which could be manipulated if the source data contains adversarial content.
Sanitization: No evidence of sanitization, filtering, or validation of the ingested external content was found in the skill's instructions.

internal-comms