Setup and Troubleshooting

Installation

Recommended: uv

cd skills/documentation/research/semantic-scholar-search
uv venv
source .venv/bin/activate
uv pip install -r requirements.txt

Alternative: standard venv

cd skills/documentation/research/semantic-scholar-search
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Verify

python3 scripts/semantic-scholar-search.py --help

API Key (optional but recommended)

The Semantic Scholar API is free without a key, but limited to 100 requests per 5 minutes. For heavier use, set an API key:

# Obtain a free key at https://www.semanticscholar.org/product/api#api-key
export SEMANTIC_SCHOLAR_API_KEY="your-key-here"

Or add to a .env file in the skill directory and load with dotenvx run --.

Activating the environment

Every session requires activating the venv:

source skills/documentation/research/semantic-scholar-search/.venv/bin/activate

Troubleshooting

HTTP 429 — rate limit exceeded

Without an API key, you are limited to 100 requests/5 minutes. Wait 5 minutes before retrying. Set SEMANTIC_SCHOLAR_API_KEY to raise the limit.

“Paper not found”

Check the ID format:

• Semantic Scholar ID: plain alphanumeric string (e.g. 204e3073870fae3d05bcbc2f6a8e263d9b72e776)
• arxiv ID: must be prefixed with arXiv: (e.g. arXiv:1706.03762)
• DOI: accepted as-is (e.g. 10.18653/v1/2020.acl-main.740)

If the paper is very recent (< 2 weeks), it may not yet be indexed.

“Author not found”

The author subcommand requires a Semantic Scholar author ID (a numeric string like 1794492), not a name. Locate the ID by running a search for the author’s papers and extracting authorId from the response.

Import error for semanticscholar

pip install semanticscholar

If running in a venv, ensure it is activated first.