The Agent Skills Directory

[SAFE]: The skill integrates with Sarvam AI's official API (https://api.sarvam.ai/v1), which is a legitimate vendor resource for the author 'sarvamai'.
[SAFE]: Recommended installation uses standard, official packages (sarvamai and openai) from public registries.
[SAFE]: Sensitive data management follows best practices by recommending the use of environment variables (SARVAM_API_KEY) rather than hardcoding credentials.
[SAFE]: The prompting guidelines in references/prompting.md are standard instructional examples for persona setting, few-shot learning, and structured output, with no malicious injection or bypass patterns detected.
[SAFE]: The skill uses a transparent OpenAI-compatible API structure and standard Python/JavaScript SDK patterns without obfuscation or dangerous dynamic execution.

chat