obsidian-challenge

Imported from herveDarritchon/ma-campagne-en-azeroth (.agents/skills/obsidian-challenge/SKILL.md). Install upstream with npx skills add herveDarritchon/ma-campagne-en-azeroth --skill obsidian-challenge. Copyright stays with the author.

Use the obsidian-second-brain skill. Execute /obsidian-challenge $ARGUMENTS:

The optional argument is the idea, belief, or plan to challenge. If not provided, infer the user's current position from conversation context.

1. Read _CLAUDE.md first if it exists in the vault root
2. Identify the user's current claim, plan, or assumption - either from the argument or from recent conversation
3. Extract the key premises behind that position
4. Search the vault for counter-evidence - spawn parallel subagents:
   • Decisions agent: search Key Decisions sections in project notes for past decisions that contradicted or reversed similar thinking
   • Failures agent: search dev logs, daily notes, and archives for past failures, regrets, or lessons learned related to this topic
   • Contradictions agent: search for notes where the user held the opposite position or flagged risks about this exact approach
5. Synthesize a structured "Red Team" analysis:
   • Your position: restate the claim clearly
   • Counter-evidence from your vault: cite specific notes, dates, and quotes
   • Blind spots: what the user might be ignoring based on their own history
   • Verdict: is this position consistent with past experience, or does the vault suggest caution?
6. Log the challenge in today's daily note under a Thinking section

Do not be agreeable. The entire point is to pressure-test. Cite specific vault files. If you find nothing contradictory, say so honestly - but search thoroughly first.

AI-first rule: Every note created or updated by this command MUST follow references/ai-first-rules.md - ## For future Claude preamble, rich frontmatter (type, date, tags, ai-first: true, plus type-specific fields), recency markers per external claim, mandatory [[wikilinks]] for every person/project/concept referenced, sources preserved verbatim with URLs inline, and confidence levels where applicable. The vault is for future-Claude retrieval - not human reading.

Anti-fabrication: Search exhaustively before claiming any note, person, or file is absent - false absence is the most common failure mode - and never invent facts, entities, or dates (mark unknowns as TBD). See the anti-fabrication and search-completeness hard rules in references/ai-first-rules.md.