The Agent Skills Directory

[SAFE]: The skill implements a highly structured, defensive reasoning process designed specifically to combat cognitive biases like confirmation bias and premature convergence. It enforces rigorous evidence-based documentation using the 'No source, no conclusion' rule.
[EXTERNAL_DOWNLOADS]: The skill references dependencies multi-agent-debate and report-output-discipline and suggests an installation command using npx skills add FuDesign2008/open-skills. This follows the standard pattern for resources from the author 'FuDesign2008' and is used to extend the agent's capabilities in a documented manner.
[COMMAND_EXECUTION]: The skill mentions executing commands (e.g., grep, cmd) as part of a 'Physical verification design' to validate hypotheses. These are instructions for the agent to perform standard investigative actions on the local environment to gather facts, which is consistent with the skill's purpose as a diagnostic tool.
[DATA_EXFILTRATION]: While the skill involves reading logs and code for analysis, there are no patterns indicating the unauthorized transmission of this data to external servers. Output is directed to local documentation paths (docs/reports/diagnosis/).

essence-diagnosis