The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and summarize data from several untrusted external sources, creating a significant surface for indirect prompt injection attacks.
Ingestion points: The agent is instructed to gather content from Slack messages, emails, Google Drive documents, and external press as detailed in examples/3p-updates.md, examples/company-newsletter.md, and examples/faq-answers.md.
Boundary markers: The instructions lack specific delimiters or warnings to ignore potentially malicious commands embedded within these external data sources.
Capability inventory: The agent reads from these sources to generate internal company communications (newsletters, status updates, FAQs) which are then distributed to employees and leadership, creating a path for malicious content to reach a wide internal audience.
Sanitization: There are no instructions for filtering, sanitizing, or validating the data retrieved from external sources before it is processed.
[METADATA_POISONING]: The skill contains inconsistent metadata regarding its licensing.
The YAML frontmatter in SKILL.md describes the license as "Proprietary," while the LICENSE.txt file provided in the skill package contains the full text for the Apache License, Version 2.0.

team-communication