The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The skill relies on executing the npx convex CLI tool to interact with remote deployments. This is the intended primary purpose of the skill.
[DATA_EXFILTRATION] (MEDIUM): The skill provides instructions for listing and retrieving environment variables using npx convex env list and npx convex env get. This functionality exposes potentially sensitive configuration data, such as API keys and secrets, to the agent's context. Because this is the primary purpose of the debugging skill, the severity is reduced from HIGH to MEDIUM.
[PROMPT_INJECTION] (LOW): Indirect Prompt Injection Surface. The skill ingests untrusted data from deployment logs via npx convex logs.
Ingestion points: Standard output from npx convex logs and npx convex logs --history.
Boundary markers: Absent. There are no delimiters or instructions provided to the agent to ignore instructions embedded within the logs.
Capability inventory: The agent has the capability to execute shell commands (npx, cd) as defined in the skill.
Sanitization: Absent. Log content is passed directly to the agent without filtering or escaping.

convex-logs