agentleFS
tograftograf/sidebrain/.claude/skills/research/SKILL.md

research

Skill1 starsChanged 5 months agoMITView on GitHub

Deep, multi-note research across the vault. Decompose a question into several searches, gather and read many related notes (following wikilinks and backlinks), then write a structured, cited analysis note into the fitting resource topic. Heavier than query, which gives a concise one-shot answer — use for "compare X and Y", "write an analysis of Z", "deep dive on…".

189 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use daily-arxiv, exp-eval and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • LambenthanSkill

    daily-arxiv

    每日从 arXiv 拉取新论文,过滤相关性,auto-ingest 高优先级论文,检测 SOTA 更新

    Also in 5 projects that use this

  • LambenthanSkill

    exp-eval

    实验判决门:Review LLM 独立评判实验结果 → 4 种判决路径 → 自动更新 claims confidence、ideas status、graph edges

    Also in 5 projects that use this

  • truongtv22Skill

    Problem-Solving Techniques

    Apply systematic problem-solving techniques for complexity spirals (simplification cascades), innovation blocks (collision-zone thinking), recurring patterns (meta-pattern recognition), assumption constraints (inversion exercise), scale uncertainty (scale game), and dispatch when stuck. Techniques derived from Microsoft Amplifier project patterns adapted for immediate application.

    Also in 5 projects that use this

  • liufeng1976Skill

    setup-ts-deep-modules

    Wire dependency-cruiser into a TypeScript repo so each package is a deep module — implementation hidden in subfolders, reachable only through its entry-point files. User-invoked.

    Also in 43 projects that use this

  • skyllwtSkill

    exp-pilot-eval

    Pilot result evaluation — read pilot results, apply success criteria, update idea page (pilot_result, failure_reason if failed), generate PILOT_VERDICT_REPORT. Called by /ideate Phase 5 after /exp-pilot-run.

    Also in 4 projects that use this

  • skyllwtSkill

    exp-pilot-run

    Pilot experiment execution — read Pilot Spec YAML, write pilot code, run experiment(Confirm with the user before operation and require the applicant to conduct manual inspection), return results. Called by /ideate Phase 5. Does NOT modify wiki pages or judge pass/fail.

    Also in 4 projects that use this

  • +5 more from Lambenthan
  • +1 more from skyllwt

The file

Files
SKILL.md
<!-- Adapted from the research skill of a private LLM-wiki repo (see BACKLOG.md), following the LLM-wiki pattern (karpathy). Reworked for this PARA vault: qmd optional, Grep fallback. -->

research — deep multi-note study, written back into the vault

research vs query: query answers one question concisely from the top notes. research deliberately sweeps broadly and produces a persisted, structured analysis note. Vault conventions live in CLAUDE.md.

Folder paths in this skill are the template defaults — the authoritative paths live in the "Folder roles" table in CLAUDE.md's Vault structure section. Look folders up there by role — never trust the literal defaults on a renamed or non-English vault.

Steps

  1. Scope & confirm. Restate the research question, the angle, the intended sections/depth, and where the analysis note will land (usually 04-Resources/<Topic>/; the relevant project file's folder when it's project-scoped; ask if there's no obvious home). Because this sweeps wide and writes a file, confirm before the heavy retrieval — this is the same gate the brainstorming skill uses.
  2. Broad retrieval. Decompose the question into sub-questions and run several searches — never rely on one query. Probe command -v qmd once:
    • qmd installed: qmd query "<sub-question>" -n 15 per sub-question.
    • No qmd: Grep key terms and synonyms across **/*.md; Glob the relevant topic folders; scan the ## Index sections of the topic starter notes. Then widen beyond direct hits: follow [[wikilinks]] outward AND backlinks (Grep for the central notes' names inside [[…]]) so related material isn't missed.
  3. Read in full. Fetch the gathered notes whole — snippets are only leads; synthesize from full text, including cited source notes.
  4. Synthesize a structured analysis:
    • A clear question/thesis up front.
    • Sections by sub-topic, every claim carrying a [[wikilink]] citation. Separate what a source claims from established fact.
    • An explicit Open questions / gaps subsection and a Contradictions subsection (lint-minded: note where notes disagree or evidence is thin).
    • A Sources list of the notes drawn on.
  5. Write the note to 04-Resources/<Topic>/Analysis - <Readable Name>.md with frontmatter (tags: [resource, analysis], created). Analysis notes are the allowed exception to one-note-one-thought atomicity — they are synthetic and longer by design.
  6. Cross-link & index. Link the analysis to the notes it draws on (they gain a backlink automatically), add it to the ## Index section of the topic starter note, and log one line in today's daily note. Follow-up todos obey the two-place todo rule.
  7. Re-embed if qmd is installed; skip silently otherwise.

Quality bar

  • Cite every claim; never assert un-sourced statements as established. Label model knowledge added on top of vault content as such.
  • Surface gaps and contradictions rather than papering over them — where the vault is thin on a sub-topic, say so and recommend an ingest to fill it.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.