본문으로 건너뛰기
cccskills
무료GitHub에 공개

Wiki Explore

Review and verify unverified wiki pages — "review unverified pages", "verify wiki content", verification pass, after a major ingest cycle, before content release. The gate that turns AI-generated knowledge into trusted knowledge.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

설치 방법 보기

포함된 파일(1)

  • SKILL.md5.3 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

Wiki Explore

Purpose

Present unverified wiki pages (explored: false) for human review and confirmation. The verification gate that turns AI-generated knowledge into trusted wiki content.

The wiki self-expands via wiki-ingest, but Maxim never publishes content that depends on explored: false pages alone. wiki-explore is the loop that closes that gap.

When To Invoke

  • User wants to review AI-generated pages
  • After a major wiki-ingest cycle (especially nightly batches)
  • Before any major content release
  • Periodic verification pass (suggested: every 1-2 weeks)

Inputs

  • filter (optional):
    • topic: only review pages matching this topic
    • since: only review pages created/updated since this date
    • startup: only review pages tagged with this startup
    • room: scope to a wing/room (sources/concepts/entities/syntheses)

Workflow

  1. Find all explored: false pages via MemPalace search.
  2. Sort: most recently updated first (newest unverified content has highest review value).
  3. Apply filters if provided.
  4. For each page: a. Display the page's TLDR b. Display the Counter-Arguments section c. Display the cited Sources d. Ask the user: Confirm | Edit | Skip | Delete e. On Confirm: set explored: true, update updated date, leave confidence as-is f. On Edit: apply user's corrections inline, then set explored: true + bump updated g. On Skip: leave explored: false, move to next page (will appear next session) h. On Delete: remove page from MemPalace; record deletion in log-md
  5. Update related: if confirmed page becomes the canonical version of a concept that had duplicates, suggest merge to other pages.
  6. Report progress at the end.

Interactive Display Format (per page)

[Page 3 of 47] Last updated: 2026-04-15

Title: Multi-Agent Handoff Patterns
Tags: [agents, architecture, governance]
Confidence: medium
Sources: raw/sources/handoff-paper.pdf, raw/clippings/multi-agent-essay.md

TLDR:
Inter-agent handoff requires explicit state transfer with status flags
(BLOCKED/PARTIAL/READY) to prevent silent failures. Maxim implements this
via .mxm-skills/agents-handoff.md.

Counter-Arguments:
- Single-agent systems avoid handoff complexity entirely (Hermes-agent thesis)
- Static handoff via files vs runtime message bus — tradeoff unresolved

Action: [C]onfirm | [E]dit | [S]kip | [D]elete | [Q]uit?

Final Report Format

Wiki Explore Session — [YYYY-MM-DD HH:MM]
Pages reviewed: [N]
Confirmed: [N]
Edited & confirmed: [N]
Skipped: [N] (will appear next session)
Deleted: [N]
Remaining unverified: [N]

Confidence: 🟢 HIGH (full pass complete) | 🟡 MEDIUM (filtered subset) | 🔴 LOW (early exit)

Suggested next:
- Run /mxm-wiki query <topic> to test the now-verified pages
- Run /mxm-wiki lint to recheck the wiki for any new contradictions

MemPalace Integration

# Find unverified pages
mempalace_search({
  query: "explored: false",
  wing: "wing_knowledge",
  limit: 50
})

# Update on Confirm
# (since MemPalace drawers are append-only, "update" = add a new drawer with same path,
# or use a maintenance script to rewrite frontmatter — depends on MemPalace version)

# Delete on Delete
mempalace_delete_drawer({ drawer_id: "<id>" })

Maxim Behavioral Layer

  • Confidence tagging: 🟢 HIGH on full pass | 🟡 MEDIUM on filtered subset | 🔴 LOW on early exit
  • Anti-rubber-stamp: every page must be SHOWN to the user (TLDR + Counter-Arguments + Sources). No bulk-confirm.
  • Trust ladder: confirmed pages move from explored: false to explored: true, unlocking citation by wiki-query

Why This Skill Exists

Without wiki-explore, AI-generated content silently propagates as "knowledge." That's how hallucination compounds. The explored-false default + this skill is the structural fix.

Handoff

  • Confirmed pages → unlock for wiki-query citation
  • Edited pages → notify wiki-lint to re-check the wiki for new contradictions
  • Deleted pages → notify wiki-ingest to re-trigger if the source remains in raw/
  • Patterns of consistent edits → suggest improvements to wiki-ingest prompts (skill-synthesizer can absorb)

Source References

  • FINAL_RELEASE_CHANGES/Brandvoice-wiki-example.md — wiki-explore pattern
  • documents/reference/AGENTS.md — wiki trust levels
  • .claude/skills/wiki-ingest/SKILL.md — produces the explored: false pages this skill verifies

Copyright (c) 2026 iSystematic Inc. Maxim product. BSL 1.1.

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

AI video and image prompt engineering — video generation, image generation, product video, social media video, cinematic direction, animation, Seedance, Higgsfield, GPT-Image-2. Applies 15 Maxim-authored cinematic styles and the atomic prompt schema.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

DrNabeelKhan/maxim22026년 10월 16일 업데이트

Apply a named behavioral science framework to any task — persuasion, nudge design, habit formation, conversion, influence, negotiation, cognitive bias, choice architecture, decision framing. The framework backbone every other Maxim skill loops into for psychological validation.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

DrNabeelKhan/maxim22026년 10월 16일 업데이트

Founder and CEO automation — morning standup, overnight cycle, burn rate, runway, pipeline review, investor and fundraising prep, executive summary, daily digest. Operates on the .mxm-executive-summary context.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

DrNabeelKhan/maxim22026년 10월 16일 업데이트

Tear down a company on request — "do a teardown of", "break down this company", "analyze this competitor's business". Fixed-template output: business model, how they make money, positioning, and where they are exposed.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

DrNabeelKhan/maxim22026년 10월 16일 업데이트

Track what a competitor actually shipped — "watch this competitor", "track what they ship", "what changed at". Compares against the last snapshot and reports only real moves: launches, pricing changes, positioning shifts.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

DrNabeelKhan/maxim22026년 10월 16일 업데이트

Regulatory review across 14 frameworks — compliance review, privacy impact assessment, regulatory gap analysis, data subject rights, consent mechanisms, audit trail, AI ethics review, data flow mapping. Covers GDPR, PIPEDA, HIPAA, PCI-DSS, SOC2, UAE-PDPL, EU AI Act and more.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

DrNabeelKhan/maxim22026년 10월 16일 업데이트

DrNabeelKhan의 스킬 모두 보기

이 스킬의 문제 신고