The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains directives that override standard agent behavior, such as 'Do not summarise the skill' and 'Do not explain what it does'. While these are intended to streamline the user experience, they represent a deliberate override of the agent's default communication style.
[INDIRECT_PROMPT_INJECTION]: The skill accepts 'context dumps' (notes, transcripts, or bullet points) from users to generate content. This creates a vulnerability surface where malicious instructions hidden within those external data sources could potentially influence the agent's logic or output during the post-writing process.

post-writer