The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on vexor, a non-standard command-line utility for semantic search. It references an external installation guide references/install-vexor.md, which involves executing unverified software from the community.
[DATA_EXFILTRATION]: The instructions for vexor doctor and vexor config indicate that the tool requires connectivity to an external API. This implies that file contents or metadata are sent to a remote service for semantic indexing and search, posing a risk of repository data exposure to an unknown destination.
[PROMPT_INJECTION]: The skill provides an attack surface for indirect prompt injection by retrieving and processing content from files within the local repository.
Ingestion points: Snippets of file content retrieved by the vexor search command and returned to the agent context.
Boundary markers: There are no markers or instructions to isolate the retrieved content from the agent's core instructions.
Capability inventory: The skill allows for shell command execution of the vexor tool.
Sanitization: The skill lacks sanitization or filtering to prevent the agent from obeying malicious instructions embedded in the files it searches.

vexor-cli