Skip to content

po4yka/llm-wiki-skills / skills/wiki-query

wiki-query

Verified

Answer questions from an existing LLM-Wiki vault and save reusable answers back into the wiki. Use when the user asks for research, synthesis, comparison, decision support, or recall from a raw/wiki/schema knowledge base.

Wiki Query

Goal

Answer from the compiled wiki first, verify against raw sources when needed, and save reusable answers back into the wiki.

When to use

  • The user asks a question answerable from the existing vault.
  • Research, synthesis, comparison or decision support should reuse wiki knowledge.
  • A reusable answer should be filed back into the wiki.

A filled question-to-answer example ships in assets/worked-example.md.

Inputs

  • User question.
  • wiki/index.md.
  • Relevant wiki pages.
  • Raw sources when verification is required.
  • Optional output format: memo, table, checklist, PRD, slides outline, report.

Write permissions

Allowed:

  • read all wiki files;
  • read raw sources when needed;
  • create wiki/queries/* for reusable answers;
  • create draft wiki/synthesis/* only when requested or clearly useful;
  • append to wiki/log.md when saving an answer.

Not allowed unless explicitly requested:

  • rewrite existing source/entity/concept pages;
  • mark claims verified;
  • edit protected human-owned sections;
  • delete or archive pages.

Procedure

1. Clarify the question internally

Identify the answer type, time sensitivity, verification needs and whether the answer should be saved. Do not ask a clarifying question when the best effort is obvious.

2. Start with the map

Read wiki/index.md first and identify candidate pages.

3. Search the wiki

Use exact search for user terms, synonyms, entity names, wikilink targets, frontmatter types and tags. Prefer focused reads over loading many full files.

4. Build an evidence set

For each important claim, classify support:

extracted | inferred | ambiguous | synthesis | unsupported | conflicting

Use raw sources when the claim is high-impact, current, legal/financial/medical/security-relevant, or the wiki page is stale.

5. Answer clearly

Distinguish:

  • what the wiki says;
  • what is directly sourced;
  • what is inferred;
  • what remains uncertain;
  • what should be checked next.

6. File back reusable output

If the answer is durable, create:

wiki/queries/YYYY-MM-DD-question-slug.md

Suggested structure:

---
type: query
status: draft
created: YYYY-MM-DD
source_paths: []
review_required: true
---

# Question

## Answer

## Evidence used

## Inferences

## Open questions

## Pages to update

Append to wiki/log.md if a page is saved.

Output

A direct answer to the user's question: a concise thesis, evidence grouped by theme and labeled extracted | inferred | ambiguous | synthesis | unsupported | conflicting, explicit uncertainty, and links to the wiki pages used. When the answer is durable, the same content is also filed as wiki/queries/YYYY-MM-DD-question-slug.md (using the structure above) with a matching entry appended to wiki/log.md.

Answer style

Prefer a concise thesis first, evidence grouped by theme, explicit uncertainty, links to relevant wiki pages and one action-oriented next step when appropriate.

Avoid pretending the wiki knows more than it does, citing generated pages as if they were raw sources, or leaving reusable answers in chat only.

Safety gates

  • Default write scope is limited to new wiki/queries/* files and, when clearly useful or requested, draft wiki/synthesis/* pages.
  • Never rewrite, delete, or archive an existing source/entity/concept page, and never mark a claim verified, without an explicit user request and review.
  • Leave protected human-owned sections untouched.
  • Every saved query page carries review_required: true in its frontmatter, and destructive edits to existing pages require explicit request and review before they happen.
add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks.Stars: 115kUpdated Claude CodeCodexVerifiedInstall