wiki-research-research
Daily research agent for
What this skill does
# Wiki Research: Research Process `#research` items from the queue — papers, academic concepts, authors, ideas worth deep-diving. Uses the 5-step loop with stricter source-finding than engineering. ## Vault location `/app/data/vault/` ## When to skip the loop If the topic is a single paper you only need to summarise, do one `web_search` (engine `brave`) for the paper, run `intelli_extract` on the abstract page, write a one-paragraph page with citation, stop. The full loop is for concepts, debates, bodies of work. ## The 5-step loop Each step has a gate. Do not advance until the gate is met. ### 1 — Define Before searching, write three sentences: - **Assumption:** what you currently believe is true about this concept / paper / author - **Unknown:** the specific thing you don't know yet, as a question with a verifiable answer - **Answer shape:** what a good answer looks like (a summary, a comparison with prior work, a critique, a position in a debate) **Gate:** a reader could tell whether your output answered it. ### 2 — Triage Identify the 2–4 source types most likely to answer. For research topics, priority order: 1. The paper itself (arXiv, publisher page, author PDF) — use `web_search` to find it, then `intelli_extract` on the abstract / introduction / results 2. Author's own page, lab page, talks, blog 3. Peer papers — critiques, extensions, follow-on work (search `<paper title>` + "follow-up" / "critique" / "reply") 4. Surveys / textbooks that cite it (search `<concept>` + "survey" or "lecture notes") 5. Adjacent vault pages (`wiki/index.md`) Run 3–4 Brave queries. For each promising URL, use `intelli_extract` with a `focusPrompt` naming the specific Unknown. **Gate:** you can name 3+ source URLs you will cite before writing the page. ### 3 — Query Pull the answer from the triaged sources. For research topics specifically: - For a paper: capture abstract, key contributions, methodology, results, reception (citations, awards), criticisms - For an author: capture background, key work, current focus, collaborators, intellectual lineage - For a concept: capture definition, origin, formal statement, what it solves, what it doesn't solve, related concepts **Gate:** every claim has at least one source URL attached. ### 4 — Verify For every claim, assign one tier: - **FACT** — directly stated in a source, with citation - **LIKELY** — consistent with sources but inferred or synthesised - **SPECULATIVE** — your reasoning beyond what sources support - **UNKNOWN** — neither you nor the sources say For any claim about the paper's reception or impact, run a second `web_search` to corroborate. Citation counts, awards, and critiques are easy to get wrong. **Gate:** every claim has a tier. ### 5 — Report Write the wiki page. Append a Limitations section and an Open Questions section. ## Wiki page format ```markdown --- tags: [#research] last_researched: YYYY-MM-DD --- # <Title> > One-sentence summary. ## Define - **Assumption:** ... - **Unknown:** ... - **Answer shape:** ... ## Findings ### Background - **FACT:** <claim> [^1] ### Key contributions - **FACT:** <claim> [^1] - **FACT:** <claim> [^2] ### Reception / critiques - **LIKELY:** <claim> [^2, 3] ### Current status - **LIKELY:** <claim> ## Open Questions - Things the research surfaced but did not resolve - Productive follow-up directions ## Limitations - What the sources did not cover - Which claims were not independently verified - What would change the answer ## Sources [^1]: <url> — accessed YYYY-MM-DD [^2]: <url> — accessed YYYY-MM-DD ``` ## Per-run workflow 1. Read `RESEARCH_QUEUE.md` in the vault root. 2. Find all unchecked items (`- [ ]`) under `## #research`. If none (or only placeholder), stop. 3. For each real unchecked item, run the 5-step loop. 4. Check `wiki/index.md` for an existing page. Update if found; otherwise create `wiki/research/<slug>.md`. 5. Update `wiki/index.md`. 6. Mark done in `RESEARCH_QUEUE.md`: `- [x]` + ` — researched YYYY-MM-DD`. 7. Move `[x]` items to `## Archive`. 8. Append entry to `wiki/log.md`: ``` ## [YYYY-MM-DD] ingest | research-queue #research | <topics> → <pages> ``` ## Rules - Never edit `raw/`, `Openclaw articles/`, or `personal stuff/`. - Only write to `wiki/` and `RESEARCH_QUEUE.md`. - Include an Open Questions section — what the research surfaced but didn't resolve. - If a topic is too vague, mark `[x]` with `— skipped: too vague, clarify topic`. - Today's date comes from system context.
Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.