The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local system commands to manage version control for the vault.
Evidence: commands/tidy.md includes instructions to execute git init, git add -A, and git commit to create a baseline before performing directory reorganization.
[DATA_EXPOSURE]: The skill references specific filesystem paths and accesses configuration files outside the immediate vault directory.
Evidence: commands/refine.md contains a hardcoded absolute fallback path /Users/hansonmei/OneDrive/obsidian-vault/ which discloses personal directory structures.
Evidence: commands/vault.md and commands/init.md read and write to a global configuration file located at ~/.proma/agent-workspaces/obsidian/obos-config.json.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-generated note content from the vault to synthesize answers and generate new content, creating an attack surface for indirect prompt injection.
Ingestion points: commands/ask.md, commands/draft.md, and commands/refine.md read the contents of Markdown files from various vault directories (Notes/, Clippings/, References/).
Boundary markers: The skill lacks explicit boundary markers or instructions to the agent to disregard instructions embedded within the notes during synthesis tasks.
Capability inventory: The skill possesses file writing capabilities across most commands and shell execution capabilities in commands/tidy.md.
Sanitization: No sanitization or filtering logic is defined for the content extracted from the Obsidian notes.

obos