The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple sources which can be manipulated to influence agent behavior.
Ingestion points: Git repositories, workspace files (including .lev, docs/, and README.md), session history (via cass), and local diagrams (~/.agents/diagrams).
Boundary markers: No specific delimiters or instructions to ignore embedded commands are used when processing retrieved content.
Capability inventory: The skill can execute shell commands (lev find, cass), perform file system operations (Grep, Glob, Read), and conduct semantic searches.
Sanitization: No evidence of sanitization or filtering of the retrieved content before it is processed by the agent.
[COMMAND_EXECUTION]: The skill executes local command-line interface tools to perform its search and retrieval functions.
Evidence: The instructions invoke cass search, cass health, cass expand, and lev find to retrieve session history and fused search results from the local environment.

prior-art