The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads project source code by cloning the repository at https://github.com/Jamkris/everything-gemini-code.git to a temporary directory. This is used as the source for the installation process.
[COMMAND_EXECUTION]: The skill executes standard filesystem commands to manage the installation, including rm for cleanup, mkdir to create configuration directories, and cp to install the selected skills and rules to ~/.gemini/ or .gemini/ directories. It also uses ls and grep to verify the installation and check for path consistency.
[DATA_EXPOSURE]: The skill reads and scans files within the user's ~/.gemini/ and local .gemini/ directories. This activity is focused on verifying that the installed skills and rules are correctly configured and identifying any broken internal references.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data by installing markdown instructions from a remote source and editing them based on user input.
Ingestion points: Content is ingested from the Jamkris/everything-gemini-code repository (SKILL.md) and via user responses to AskUserQuestion (SKILL.md).
Boundary markers: No explicit delimiter or 'ignore' instructions are used when interpolating user choices into file edits.
Capability inventory: The skill has capabilities for shell command execution (rm, git clone, mkdir, cp, ls, grep) and in-place file editing (SKILL.md).
Sanitization: The skill modifies local .md files based on user-provided tech stack preferences and optimization choices.

configure-ecc