The Agent Skills Directory

[PROMPT_INJECTION]: The skill defines a specific analytical persona and response style. It includes instructions to decline tasks outside its scope (such as coding or factual research) and provides guidelines to avoid typical AI speech patterns. These are legitimate scoping and stylistic instructions rather than attempts to bypass core safety guardrails.- [COMMAND_EXECUTION]: There are no shell scripts, system commands, or executable files within the skill package. The instructions do not invoke any platform tools or command-line interfaces.- [EXTERNAL_DOWNLOADS]: The skill does not contain any URLs for remote code, nor does it reference external package managers like npm or pip. All reference materials are stored locally as Markdown files.- [DATA_EXFILTRATION]: No network operations or data harvesting patterns were detected. The skill operates purely within the conversational context provided by the agent.- [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or credentials were found in the provided files. Instruction for handling secrets is not applicable as the skill does not perform authenticated operations.

zizek