The Agent Skills Directory

[SAFE]: The skill is purely instructional and contains no executable code, scripts, or binaries. It functions as a methodology guide for prompt engineering.
[SAFE]: All external URLs point to legitimate academic, research, and educational repositories including arXiv, official technology company blogs (e.g., Anthropic, Palantir), and established AI safety/learning platforms. No suspicious or obfuscated links were detected.
[SAFE]: The skill includes explicit ethical boundaries in its flowchart and 'When NOT to Use' sections, prohibiting the use of its techniques for deception, fraud, or harassment.
[SAFE]: The data processing logic (translating metrics like Flesch-Kincaid grades or function-word rates into prompt constraints) follows standard industry practices for persona-based prompting and does not involve unsafe operations such as shell execution or sensitive file access.

subagent-instruction-operationalization