The Agent Skills Directory

[SAFE]: The skill is a collection of Markdown-based guidelines and agent instructions focused on prompt engineering best practices.
[COMMAND_EXECUTION]: The SKILL.md file provides installation commands (git clone, make install) for the user to follow manually. There are no instructions for the agent to execute arbitrary shell commands or remote scripts during operation.
[DATA_EXPOSURE]: The agent is instructed to read from local paths such as references/ and ../_shared/ to retrieve policy and pattern documentation. These paths are internal to the skill's deployment and do not involve access to sensitive system directories or user credentials.
[INDIRECT_PROMPT_INJECTION]: While the skill's primary purpose is to process and generate user-supplied prompts, it does not possess dangerous capabilities like network access or file-writing tools that could be leveraged by a malicious input. The instructions emphasize maintaining a clear 'commander's intent' and using recognition tests, which are good practices for reducing the impact of adversarial inputs.

prompt-authoring