The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the curation of user-authored notes by reading and analyzing their content to suggest links and metadata. This introduces a surface for indirect prompt injection if a note contains malicious instructions intended to manipulate the agent's behavior during curation.
Ingestion points: The agent reads note content via read_note and search_notes as described in SKILL.md.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when processing note data.
Capability inventory: The agent is equipped with file-system and note-management capabilities including edit_note, write_note, and move_note (SKILL.md).
Sanitization: No sanitization or escaping mechanisms for note content are mentioned prior to the analysis of the knowledge graph.
[COMMAND_EXECUTION]: The skill refers to a command-line tool, bm (e.g., bm orphans), which is a vendor-provided utility for identifying unlinked notes. This usage is consistent with the skill's primary purpose of knowledge base management.

memory-curate