The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's analyze command parses data from untrusted files in the current repository (e.g., README.md, .cursorrules, and other agent configs) to synthesize persistent notes in an Obsidian vault. These notes are automatically retrieved and read during the 'Session Start' phase to orient the agent, creating a pathway for malicious instructions embedded in a project to influence the agent's future behavior.
Ingestion points: The analyze command fully reads documentation and agent configuration files from the project directory.
Boundary markers: The skill does not define specific delimiters or instructions for the agent to ignore or isolate instructions found within the synthesized vault content.
Capability inventory: The skill can execute shell commands (git, obsidian CLI, filesystem operations), write to the persistent vault, and read vault content automatically upon session start.
Sanitization: No sanitization or validation of the ingested documentation content is performed before it is formatted into vault notes.
[COMMAND_EXECUTION]: The skill extensively uses shell command execution to function. This includes git for repository metadata, standard filesystem operations (mkdir, cp), and an external obsidian CLI tool for managing properties and links within the knowledge vault.

obs-memory