The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes Obsidian-flavored Markdown (.md), JSON Canvas (.canvas), and Obsidian Bases (.base) files. These serve as ingestion points for untrusted data that could contain malicious instructions for the agent.
Ingestion points: Operations described in SKILL.md for reading and editing note files, including properties, callouts, and canvas nodes.
Boundary markers: The skill instructions do not specify any delimiters or safety markers to differentiate between note content and agent instructions.
Capability inventory: The skill allows the agent to perform file system writes and modifications to structured files.
Sanitization: There is no evidence of sanitization or filtering logic to prevent the agent from obeying instructions embedded within the processed notes.
[COMMAND_EXECUTION]: The SKILL.md file contains a section titled "后续维护" (Follow-up maintenance) that provides a PowerShell code block using Invoke-WebRequest and Expand-Archive. While intended for manual updates, these commands could be executed by an agent attempting to perform maintenance.
[EXTERNAL_DOWNLOADS]: The skill provides instructions to download a ZIP archive from a repository hosted on GitHub (github.com/kepano/obsidian-skills). The download targets a well-known service and is documented as a maintenance step.

obsidian-skills-integrated