The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions include a "self-healing" mechanism where the agent is prompted to update its own SKILL.md and registry files when new context is discovered. This design feature allows the agent to modify its own instructions based on processed data.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface through the ingestion of user-provided content for note creation.
Ingestion points: The body and body_file arguments in scripts/create_branium_note.py.
Boundary markers: Absent; note content is written directly without delimiters.
Capability inventory: The script performs file writes using Path.write_text.
Sanitization: Note titles are sanitized for filename safety, but the body content is not validated or escaped.
[COMMAND_EXECUTION]: The test suite utilizes subprocess.run to execute the local Python script for functional testing. This is a standard development practice and does not involve risky shell execution or remote code.
[SAFE]: The skill uses hardcoded absolute paths to a specific local user directory. While this exposes the folder structure of the target system, it is necessary for the skill's primary purpose of vault management.

document-branium