mode-02-tool-calling-accuracy

SKILLWorkflowCommunity
v0.0.0berto-playUnknownAktualisiert vor 22 TQuelle →

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

Community-submitted skill. Not yet reviewed by the Forge team. Full prompt content may not be available.Request review →
1Clients
1Formate
vor 22 TLetzte Aktualisierung
Skill
Autorberto-play
Version0.0.0
LizenzUnknown
KategorieWorkflow
Formateskill.md
PromptNicht veröffentlicht
Kompatibilität
Claude✓ Unterstützt
Cursor
Copilot
ChatGPT
Gemini
Über

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 an L0 to L4 pipeline: justify the tool against intent, ground every parameter in a named source (nev

Schlagwörter
skillclaude