The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions in SKILL.md include a functional command template for tutorial writing that downloads and executes a shell script from a remote domain (https://opencode.ai/install) by piping it directly to bash. Executing remote scripts from unverified domains is a high-risk activity that can lead to system compromise.
[COMMAND_EXECUTION]: The skill encourages the inclusion of shell commands in technical content, specifically demonstrating the use of curl and bash for software installation without providing warnings or verification steps for the source code.
[INDIRECT_PROMPT_INJECTION]: The skill exposes a surface for indirect prompt injection by ingesting and processing untrusted user text (/简写, /风格检查) for style modification without implementing boundary markers or instructions to ignore embedded directives.
Ingestion points: User-supplied topics and text provided to the writing and style checking commands in SKILL.md.
Boundary markers: Absent. There are no delimiters or instructions provided to the agent to treat user input as data rather than instructions.
Capability inventory: The skill provides templates for executing remote shell scripts via network operations.
Sanitization: Absent. No escaping or validation of user-provided content is mentioned in the instructions.

brief-write