The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The scripts/setup.sh script downloads an environment initialization script from the developer's official GitHub repository at github.com/FinpeakInc/frevana-cli-releases. This is a vendor-owned resource used for legitimate setup purposes.
[REMOTE_CODE_EXECUTION]: The skill executes the downloaded setup script using exec bash. This is a declared part of the initialization process for the Frevana service and originates from a trusted vendor source.
[COMMAND_EXECUTION]: The scripts/ask_chatgpt.sh script executes shell commands to interact with the frevana CLI tool and invokes python3 with embedded code to perform JSON serialization of data.
[PROMPT_INJECTION]: The skill is designed to process user-provided prompts and send them to an external LLM.
Ingestion points: User prompts are accepted via the --prompt argument or through a file specified by --prompt-file in scripts/ask_chatgpt.sh.
Boundary markers: No explicit delimiters or safety instructions are added to wrap the user content before it is passed to the tool.
Capability inventory: The skill has the capability to perform network requests through a local daemon and write output files.
Sanitization: Input is safely passed to a Python subprocess via environment variables and JSON-encoded, which effectively prevents command injection at the execution layer.

chatgpt-ask