The Agent Skills Directory

[PROMPT_INJECTION]: The skill interpolates user-provided text directly into an LLM prompt in src/index.ts.
Ingestion points: The idea argument provided via the CLI in src/cli.ts.
Boundary markers: None present; user input is appended directly to the end of the prompt.
Capability inventory: The skill has the capability to write files to the local file system.
Sanitization: None present; the input is passed directly to the OpenAI API.
[COMMAND_EXECUTION]: The CLI allows users to specify an output file path via the -o or --output flag in src/cli.ts. The tool uses fs.writeFileSync to write generated content to this path. While there is no path validation (allowing for potential overwriting of system files if misused), this is a standard functional feature of a CLI utility and does not appear to be malicious.
[SAFE]: All identified dependencies (commander, openai, ora) are standard, reputable packages sourced from the official NPM registry. No suspicious network activity or code obfuscation was detected.

pitch-gen