The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a local binary named evernote to perform operations. The execution is limited to specific commands (status, authorize-url, list-tools, call-tool) defined in the manifest, and write operations require explicit user approval by default.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it retrieves and processes content from external Evernote notes which could contain malicious instructions.
Ingestion points: Note content and library data (notebooks, tags) are ingested into the agent context via the notes.read and library.read methods.
Boundary markers: The current instructions do not specify the use of delimiters or clear markers to separate retrieved note content from the agent's system instructions.
Capability inventory: The skill can perform network operations (via the Evernote API) and write data to the user's account through the notes.create method.
Sanitization: The instructions do not explicitly describe sanitization or filtering logic for the content retrieved from Evernote.

evernote