The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a shell script (scripts/anki.sh) to manage flashcards, which involves executing system commands like curl to interact with the local AnkiConnect service and jq for processing JSON data.
[INDIRECT_PROMPT_INJECTION]: The skill has a potential surface for indirect prompt injection because it retrieves and processes content from the user's flashcards, which could contain instructions intended to influence the agent's behavior.
Ingestion points: The skill fetches card content, note information, and due cards from the local AnkiConnect API via the find, info, due, and review actions in scripts/anki.sh.
Boundary markers: Instructions in SKILL.md emphasize a mandatory 'CRITICAL: User Approval Required' step, directing the agent to always seek user confirmation and display all content before proceeding with modifications.
Capability inventory: The skill can add, update, and delete flashcards, as well as trigger synchronization with AnkiWeb, providing a significant capability tier for data modification.
Sanitization: There is no specific evidence of sanitization or validation logic applied to the flashcard data within the anki.sh script to filter out potentially malicious prompt content.

anki