The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute bash commands that interpolate variables such as <title>, <note>, and <content> directly into shell strings. Examples include obsidian create path="..." content="<note>" and obsidian append ... content="...". If the synthesized note content or title contains shell metacharacters like backticks (`), semicolons (;), or command substitutions ($()), it could result in arbitrary command execution on the host system.
[INDIRECT_PROMPT_INJECTION]: The skill processes summaries and insights derived from the current conversation, which may contain instructions or data from untrusted external sources (e.g., web pages or documents analyzed by the agent).
Ingestion points: Current conversation synthesis, insights, and technical decisions (SKILL.md).
Boundary markers: None identified in the shell command templates or the instructions.
Capability inventory: File creation, modification (append), and searching via the obsidian and qmd CLI tools (SKILL.md).
Sanitization: None; there are no instructions to validate, filter, or escape the content before it is used in shell commands.

save-note