From llmwiki
Detect contradictions ("needs review" flags) in .llmwiki/ and seek human judgment for resolution.
How this skill is triggered — by the user, by Claude, or both
Slash command
/llmwiki:metabolizeThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Detect contradictions ("needs review" flags) in .llmwiki/ and seek human judgment for resolution.
Detect contradictions ("needs review" flags) in .llmwiki/ and seek human judgment for resolution.
Research by DeltaZero has proven that LLM accuracy degrades exponentially with accumulated contradiction volume delta, following S = mu x e^(-delta x k). Unresolved contradictions in the wiki degrade the accuracy of all LLM operations that reference it.
If run following /llmwiki:lint, /tmp/llmwiki_lint.xml already exists. In that case, skip preprocessing and use the existing output as-is.
Run preprocessing only if the file does not exist:
Resolve input_dir: read from .llmwiki/config.json, or fall back to the project root (cwd).
python3 ${CLAUDE_PLUGIN_ROOT}/skills/make/scripts/llmwiki_preprocess.py <input_dir> --llmwiki-dir .llmwiki > /tmp/llmwiki_lint.xml
Retrieve contradiction pages from the <contradictions> section in /tmp/llmwiki_lint.xml.
If no pages are found, report "no contradictions" and stop.
Read each contradiction page and extract locations with "needs review" flags. Classify contradictions as follows:
Present a resolution proposal for each contradiction:
Resolve only contradictions approved by the user. When updating wiki pages:
updated to todayFor none type:
updated to todayAppend an entry to .llmwiki/log.md in the following format:
## [YYYY-MM-DD] metabolize | Resolved <count> (temporal:<n>, scope:<n>, genuine:<n>, none:<n>), remaining <n>
npx claudepluginhub ktrysmt/claude-plugins --plugin llmwikiFetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.
Applies a firm's KYC/AML rules grid to parsed onboarding records: assigns risk rating, checks required documents, outputs rule outcomes with citations, and routes for escalation.
Generates daily or weekly digests of activity from connected sources (chat, email, docs, tasks, CRM), highlighting action items, decisions, mentions, and project updates.