The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes potentially untrusted content from user-edited files to suggest new notes or updates. * Ingestion points: Target markdown files and the vault index file (Cards/_index.md). * Boundary markers: There are no explicit instructions to use delimiters or ignore instructions within the ingested text during analysis. * Capability inventory: The skill has access to Write, Edit, and Bash tools for file system modification. * Sanitization: The instructions do not define a process for sanitizing ingested text before extraction tasks. * Control: A critical safety measure is included where the agent is instructed to never commit changes without explicit user approval ('Automate the question', 'The user sees all suggestions and chooses').

atomize-zettel