The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains purely instructional content regarding microcopy and interface design principles. There are no patterns suggesting attempts to override safety guidelines, extract system prompts, or bypass agent constraints.
[DATA_EXFILTRATION]: No network-related commands (curl, wget, etc.), hardcoded credentials, or sensitive file path references were identified. The skill does not instruct the agent to transmit data externally.
[REMOTE_CODE_EXECUTION]: The skill does not reference external dependencies, package managers (npm, pip), or remote script execution patterns. It contains no executable scripts.
[OBFUSCATION]: A thorough scan for Base64 encoding, zero-width characters, homoglyphs, and hidden text patterns yielded no results. All content is clear and human-readable.
[PRIVILEGE_ESCALATION]: No commands for system-level modifications, such as sudo, chmod, or registry edits, are present.
[PERSISTENCE_MECHANISMS]: The instructions do not involve any mechanisms to persist across sessions, such as modifying startup scripts or cron jobs.
[DYNAMIC_CONTEXT_INJECTION]: The skill does not use shell execution placeholders or dynamic context injection syntax.

better-writing