This skill should be used when the user asks to "optimize prompts", "design prompt templates", "evaluate LLM outputs", "build agentic systems", "implement RAG", "create few-shot examples", "analyze token usage", or "design AI workflows". Use for prompt engineering patterns, LLM evaluation frameworks, agent architectures, and structured output design.
Security Analysis
medium confidenceThe skill's files and runtime instructions are consistent with a prompt-engineering toolkit and do not request extra credentials or installs, but one included script was truncated in the provided content so you should review it before running.
Name/description (Senior Prompt Engineer) match the provided artifacts: a prompt optimizer, a RAG evaluator, and an agent orchestrator. The declared requirements (no env vars, no binaries) are reasonable for the offline/static analysis and visualization tools included.
SKILL.md directs the agent to run local Python scripts against user-provided files (prompts, contexts, agent configs). That scope is appropriate. Caveat: the full content of scripts/rag_evaluator.py was not included in the truncated listing; if that script performs network calls (LLM APIs or remote evaluation) it may require credentials or contact external endpoints even though none are declared — review that file before executing.
No install spec is provided (instruction-only usage) and included code is Python scripts that operate on local files. No remote downloads, package installs, or archive extraction are present in the metadata.
The skill declares no required environment variables, credentials, or special config paths. The scripts shown operate on files supplied by the user and do not reference secret env vars in the visible code.
The skill is not always: true and does not request persistent system presence. It does not modify other skills or system-wide settings in the visible code.
Guidance
This package appears coherent for prompt engineering tasks: it contains local Python tools for prompt analysis, RAG evaluation, and agent visualization and does not declare any credentials or installers. Before running it: (1) open and review scripts/rag_evaluator.py (the listing was truncated) to confirm it does not call external APIs or expect hidden credentials; (2) inspect all scripts for any hard-coded endpoints or secrets; (3) run the tools on non-sensitive/test data first and in an isolated environment (virtualenv/container); and (4) be aware the YAML parser in agent_orchestrator.py is a simple custom parser (may mis-parse complex YAML). If rag_evaluator or any script needs LLM/remote APIs, that should be declared as required env vars — if you see such network calls without declared requirements, treat that as a red flag and do not run until you understand how credentials are provided.
Latest Release
v1.0.0
Initial release of Senior Prompt Engineer: a toolkit for advanced prompt engineering, LLM evaluation, and agentic workflow design. - Provides CLI tools for prompt optimization, Retrieval-Augmented Generation (RAG) evaluation, and agent workflow orchestration. - Includes workflows for prompt optimization, few-shot example design, and structured output specification. - Supports analysis of token usage, prompt clarity, and performance reporting. - Features validation and visualization of agent architectures. - Supplies quick-start commands, usage examples, and reference patterns for rapid adoption.
More by @alirezarezvani
Published by @alirezarezvani on ClawHub
