The Agent Skills Directory

[COMMAND_EXECUTION]: The script vikunja_tool.py executes joplin_tool.py from a sibling skill to perform note updates.
Evidence: Found in the _call_joplin function in scripts/vikunja_tool.py.
Context: The call uses a list of arguments and avoids a shell, which is a secure practice for local integration between skills from the same author.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes untrusted data from an external Vikunja instance and existing Joplin notes.
Ingestion points: Task data from Vikunja API and note bodies from Joplin API in vikunja_tool.py.
Boundary markers: None present.
Capability inventory: Network requests via httpx and file/note updates via subprocess.run.
Sanitization: Uses list-based subprocess arguments and standard string processing for Markdown generation.

vikunja