The Agent Skills Directory

[DYNAMIC_EXECUTION]: The scripts scripts/build_literature_canvas.py and scripts/build_literature_graph.py modify the Python system path at runtime to import a module named kb_common from a directory located two levels above the skill folder (../../obsidian-project-kb-core/scripts). This dynamic loading from a computed path represents a risk if the target directory contains untrusted code.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources, specifically literature notes likely imported via Zotero, which creates a potential surface for indirect prompt injection.
Ingestion points: The scripts read content from all markdown files within the Sources/Papers/ and Knowledge/ directories.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard potential instructions embedded within the paper notes or synthesis files.
Capability inventory: The associated scripts have the capability to read from and write to the local file system (using Path.read_text, Path.write_text, and Path.mkdir).
Sanitization: The scripts use regular expressions and basic string manipulation to extract metadata and links, but they do not perform sanitization to filter out malicious prompt instructions that could be executed by the agent when viewing the resulting canvas or graph.

obsidian-literature-workflow