The Agent Skills Directory

[SAFE]: The skill is entirely composed of Markdown documentation, historical examples, and instructional templates. It contains no executable code, scripts, or configuration files that could pose a security risk.
[PROMPT_INJECTION]: Instructions use standard prompting techniques to guide the agent through a logical game theory analysis process. There are no attempts to bypass safety filters, extract system prompts, or override core instructions. The use of "[WAIT]" markers is for conversational flow control and does not present an injection risk.
[DATA_EXPOSURE]: No hardcoded credentials, API keys, or sensitive file paths were found. The skill does not request access to the environment, file system, or private data.
[EXTERNAL_DOWNLOADS]: The skill includes links to informational resources, academic papers, and government websites (e.g., Anthropic, Future of Life Institute, UK Government, EU Law). These are used for reference and education, not for automated downloads or code execution.
[REMOTE_CODE_EXECUTION]: There are no patterns involving shell command execution, remote script fetching (e.g., curl | bash), or dynamic code evaluation.
[OBFUSCATION]: The content is clear and uses standard Markdown formatting. No hidden text, base64-encoded instructions, or homoglyph-based obfuscation techniques were detected.

prisoners-dilemma