The Agent Skills Directory

[PROMPT_INJECTION]: The skill provides structured guidelines for technical writing (Diátaxis, Google style) to improve agent output quality. It does not contain instructions to bypass safety filters, ignore system prompts, or reveal internal configuration.
[DATA_EXFILTRATION]: No sensitive data exposure or exfiltration patterns were detected. The skill does not use network tools or access sensitive files like credentials or environment variables.
[REMOTE_CODE_EXECUTION]: The skill does not download or execute external code. It refers to other skills ('unslop', 'write-pr') for workflow integration, which is standard agent behavior.
[COMMAND_EXECUTION]: There are no shell commands or scripts included in the skill. The 'disable-model-invocation: true' frontmatter field indicates the skill is intended as a set of instructions rather than an executable tool.
[OBFUSCATION]: The content is written in clear, plain text. No Base64, hex encoding, zero-width characters, or homoglyph substitutions were found.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to process user-provided documentation and PR descriptions, it lacks exploitable capabilities such as file system writes, network operations, or arbitrary command execution, mitigating the risk of data poisoning.
[DYNAMIC_CONTEXT_INJECTION]: The skill does not utilize dynamic context placeholders or command execution at load time.

technical-writing