The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions in SKILL.md direct the agent to execute 'npx --yes submit-expo-feedback@latest'. This command downloads and executes code from an unverified package on the npm registry at runtime, which is a high-risk pattern for arbitrary code execution.
[COMMAND_EXECUTION]: The skill utilizes shell command execution for its feedback reporting mechanism. The use of the '--yes' flag with npx bypasses user confirmation prompts during package installation and execution.
[EXTERNAL_DOWNLOADS]: The skill relies on fetching an external dependency (submit-expo-feedback) from a public registry without version pinning or integrity verification, exposing the agent to potential supply chain attacks.
[PROMPT_INJECTION]: The feedback submission template in SKILL.md creates a surface for indirect prompt and command injection. (1) Ingestion points: User-supplied feedback is interpolated into the '' placeholder. (2) Boundary markers: The input is wrapped in double quotes, but there are no instructions for the agent to handle or escape shell metacharacters like semicolons or backticks. (3) Capability inventory: The skill provides the capability to execute shell commands via npx. (4) Sanitization: There is no evidence of sanitization or validation of the user-provided feedback before it is passed to the shell.

expo-router