The Agent Skills Directory

[SAFE]: The skill consists of educational markdown instructions and reference materials designed to guide an agent through research evaluation. No executable code, network exfiltration, or hardcoded credentials were found.
[INDIRECT_PROMPT_INJECTION]: The skill is intended to process external scientific data, creating an attack surface for indirect prompt injection. However, the skill provides a highly structured and systematic methodology which focuses the agent's attention on specific methodological and statistical evidence, naturally reducing the likelihood of following arbitrary instructions embedded in the source text. The risk is typical for data-processing skills and no specific implementation flaws were detected.

scientific-critical-thinking