prompt engineer
PassAudited by ClawScan on May 1, 2026.
Overview
This appears to be a coherent prompt-engineering reference skill with user-directed examples; the main things to notice are OpenAI API-key use, provider calls, and an unpinned example dependency install.
This skill is reasonable to install as prompt-engineering guidance. Before running the included examples, inspect the scripts, use a limited OpenAI API key, keep .env private, pin dependencies for real projects, and do not send sensitive test data to external providers unless you intend to.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Running the generated evaluator can consume the user's OpenAI quota or incur API costs.
The scaffolded project expects the user to place an OpenAI API key in a local .env file so the evaluation script can use the user's OpenAI account.
OPENAI_API_KEY=sk-your-key-here ... cp .env.example .env # Add your API key
Use a limited-purpose API key, keep .env out of version control, and rotate the key if it is exposed.
If users run the examples with real data, prompts and test cases may be transmitted to an external AI provider.
The example evaluation framework sends test inputs, prompts, and outputs to the OpenAI API as part of prompt evaluation.
response = client.chat.completions.create(... messages=[... {"role": "user", "content": test.input} ...])Avoid using confidential, regulated, or customer data in evaluations unless the provider use is approved and covered by the user's data-handling policy.
A future package version could behave differently from the version the examples were written for.
The scaffolded next steps install the OpenAI package without a version pin, which is a minor supply-chain reproducibility note.
pip install openai
Pin and review dependency versions in real projects, for example by using a requirements file or lockfile.