Use when Shane wants to pressure-test a belief, steelman an opposing view, or invoke /challenge.
Delegate to Gemma via the stepped execution protocol. Claude orchestrates; Gemma executes.
mcp__ollama-agent__gemma_start with:
task: "Vault access (bash only, no MCP tools): obsidian search query='TERM' limit=10, obsidian read file='Note Name' (no .md). Steelman the strongest opposition to '[argument]'. Identify the most vulnerable assumptions, surface counterevidence from the vault, and pressure-test the position rigorously."skill: "challenge"context: any relevant context from the current conversationstatus is "running", call mcp__ollama-agent__gemma_continue with session_id; repeat until status is "done" or "error"result, synthesize if needed, and present to ShaneVault access (bash only, no MCP tools): obsidian search query='TERM' limit=10, obsidian read file='Note Name' (no .md).
Steelman the strongest opposition to '[argument]'. Identify the most vulnerable assumptions, surface counterevidence from the vault, and pressure-test the position rigorously.
Execute the skill directly:
obsidian search query='[argument]' limit=10 via bash to find notes related to the topicobsidian read file='[note name]' via bash for each — look especially for where Shane states the position most confidently[argument] — not a strawman, the actual best version[argument]; name them explicitly[argument] to be wrong?