The Agent Skills Directory

[SAFE]: The skill implements a standard text-based ranking algorithm (BM25) using standard Python libraries and mathematical formulas.
[SAFE]: The Python script scripts/bm25.py uses only built-in modules (math, json, argparse, re, collections) and performs all calculations locally without any network access or external dependencies.
[SAFE]: Input data is processed as structured JSON. The script includes a tokenize function that utilizes a strict alphanumeric regex ([A-Za-z0-9']+), which effectively sanitizes input by discarding most special characters and potential injection symbols.
[SAFE]: No malicious patterns such as prompt injection, data exfiltration, persistence mechanisms, or privilege escalation were identified. The instructions and reference materials are purely technical and consistent with the skill's stated purpose.

algo-ecom-bm25