The Agent Skills Directory

[COMMAND_EXECUTION]: The skill operates primarily by executing shell commands via bunx or npx to interact with the @temps-sdk/cli package. Use of these runners is confined to a specific, pinned version (0.1.38).
[EXTERNAL_DOWNLOADS]: Fetches and executes the @temps-sdk/cli package from the NPM registry. The skill implements a security control by performing an integrity check: it retrieves the package's checksum using npm view and compares it against a hardcoded SHA-512 hash before allowing execution. This practice helps ensure the artifact has not been tampered with at the registry level.
[INDIRECT_PROMPT_INJECTION]: The skill acknowledges the risk of processing untrusted data from CLI outputs, logs, and repository metadata. It includes explicit instructions to treat such data as untrusted and never execute any instructions found within them, which serves as a mitigation for indirect injection attacks.
[CREDENTIALS_SAFE]: The skill contains comprehensive safety guidelines for handling sensitive data. It instructs the agent to never print secrets in chat or logs, to use placeholders for flags that require secrets, and to prefer secure methods like interactive prompts or environment variables over command-line arguments.

temps-cli