The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads the gate-cli binary from the vendor's official GitHub repository (github.com/gate/gate-cli) during the setup process. The setup.sh script also dynamically fetches the latest release version from the GitHub API.
[COMMAND_EXECUTION]: The skill executes the gate-cli binary via subprocess calls to perform read-only operations such as listing activities, fetching entry details, and querying activity types. It implements a safety protocol of running --help before subcommands to identify and satisfy required arguments.
[PRIVILEGE_ESCALATION]: The setup.sh installation script utilizes sudo install as a fallback mechanism to install the binary into /usr/local/bin if the preferred local user directory is unavailable.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from the Gate Exchange API via the gate-cli cex activity list command, which could potentially contain malicious instructions in activity metadata.
Ingestion points: Data enters the context through gate-cli cex activity list and gate-cli cex activity get-entry as specified in SKILL.md.
Boundary markers: The skill enforces strict response templates and Markdown table structures to isolate external content.
Capability inventory: The skill is limited to read-only shell execution of the gate-cli tool.
Sanitization: SKILL.md mandates strict field mapping, translation of non-English keywords, and explicit URL processing rules to ensure data integrity.

gate-exchange-activitycenter