The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes local shell commands for environment discovery and metadata generation:
Executes git rev-parse --show-toplevel to identify the current project root.
Executes find ~ -maxdepth 6 -name ".obsidian" -type d to discover Obsidian vaults within the user's home directory.
Executes date "+%Y-%m-%d %H:%M" to retrieve accurate timestamps for workspace nodes.
[EXTERNAL_DOWNLOADS]: The skill provides an attribution link to the JSON Canvas specification repository at https://github.com/kepano/obsidian-skills/tree/main/skills/json-canvas. This is used for reference and format compliance.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes data from external files.
Ingestion points: Reads project-to-canvas mappings from ~/.config/work-tree-canvas/registry.json and work tree data from .canvas files.
Boundary markers: Uses the structured JSON Canvas format to delimit data; however, there are no specific prompt delimiters mentioned to prevent instructions inside text nodes from being interpreted as commands.
Capability inventory: Includes file read/write access, directory creation, and the execution of specific shell commands (git, find, date).
Sanitization: The skill implements strict validation against the JSON Canvas 1.0 specification before performing any write operations.

work-tree-canvas