The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions to install the TiDB Cloud CLI (ticloud) by downloading and executing a shell script directly from the vendor's official GitHub repository (https://raw.githubusercontent.com/tidbcloud/tidbcloud-cli/main/install.sh | sh). While this pattern is generally flagged as risky, it is a standard installation method for the official tool provided by the skill's author (pingcap).
[COMMAND_EXECUTION]: The skill performs extensive shell command execution using the ticloud CLI to manage clusters, projects, and branches. This includes potentially sensitive operations such as authentication, resource deletion, and user management.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data by running CLI commands like ticloud serverless list and ticloud project list, then parsing the output to render ASCII tables. This creates a potential surface for indirect prompt injection if resource names (e.g., cluster display names) were to contain malicious instructions.
Ingestion points: CLI output from ticloud command executions (e.g., listing regions, projects, and clusters) processed in SKILL.md.
Boundary markers: None explicitly defined for command outputs; data is rendered directly into ASCII tables.
Capability inventory: Subprocess execution via shell commands (ticloud), file-system access (implied by CLI config), and network operations (CLI communication with TiDB API).
Sanitization: No specific sanitization or escaping logic is described for the content being formatted into tables.
[SAFE]: The skill implements several security guardrails, including mandatory user confirmation before any delete operation and explicit instructions to never read or display stored passwords from .env files. The use of the --insecure-storage flag for authentication is a documented feature of the vendor's CLI for specific environments.

tidbx