Skill

mode--02-tool-calling-accuracy

A session-level enforcement MODE that makes tool and function calling reliable. It kills the three failures that break tool use: wrong tool chosen, wrong or hallucinated parameters, and silent failure (a call returns success but the output is wrong and the model proceeds anyway). Runs TCV-01 across

Claim this listing

Connect your GitHub to prove you own or maintain this listing. We verify repo access automatically — most publishers are confirmed in seconds.

1Connect GitHub

2Submit your claim

3Auto-verified, or reviewed within 48h