The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources, which constitutes a potential vector for indirect prompt injection.
Ingestion points: Retrieves performance events, deployment insights, function logs, tables metadata, and source code through the official Convex MCP.
Boundary markers: The instructions do not specify the use of delimiters or explicit guidance to the agent to ignore instructions embedded within the processed deployment data.
Capability inventory: The skill has the capability to read local source code and deployment metadata. It also has the ability to apply code modifications, though this is restricted by a requirement for explicit user confirmation.
Sanitization: There is no documented process for sanitizing or validating the data retrieved from the Convex deployment prior to its analysis by the agent.
[EXTERNAL_DOWNLOADS]: The skill recommends the use of the @convex-dev/sharded-counter package, which is an official component from the vendor's ecosystem.

convex-advisor