AGENT SKILL

citation-manager

Manage citations and references — collecting sources, organizing libraries, formatting bibliographies, and keeping reference integrity. Use when writing anything that cites sources.

License 许可
MIT
所属仓库
TanNhatCMS/jarvis
最近更新
2026-11-07
Stars
★ 0
Forks
1

01 前言 Front Matter

name
citation-manager
category
ai-research
description
Manage citations and references — collecting sources, organizing libraries, formatting bibliographies, and keeping reference integrity. Use when writing anything that cites sources.

03 SKILL.md 正文

Citation Manager

Citations are the paper's supply chain: every claim traceable to its source. Managing them means collecting sources systematically, organizing them so you can find anything, formatting consistently, and — critically — verifying that every citation is real and says what you claim.

Overview

The workflow: capture sources as you read (never "I'll find it later"), store them with notes on why they matter, cite-while-you-write with keys instead of hand-typed references, and generate the bibliography from the library. The integrity rule: every in-text citation matches a real source you've read, and every bibliography entry is cited. Reference managers automate formatting; they don't automate honesty.

When to use

  • Writing papers, theses, reports, or any document with references.
  • Building a personal research library over time.
  • Collaborating: shared libraries keep everyone's citations consistent.
  • Cleaning up a manuscript's references before submission.

Core concepts

  • Capture at reading time: save the full citation + PDF + your notes the moment you read something useful. "I'll relocate it later" fails reliably.
  • Cite keys: stable keys (smith2024attention) used in drafts; the formatted reference is generated, never hand-typed. Renumbering-proof.
  • Annotated library: each entry with a one-line note — what it claims, why you cited it. Future you will thank present you.
  • Style consistency: one citation style per document (APA, IEEE, Chicago…), applied by the tool, not by hand. Switching styles should be one setting.
  • Reference integrity: every citation verified — the source exists, you read the relevant part, it supports your claim. AI-generated citations must be checked individually.
  • Deduplication: the same source saved three ways creates triple entries. Merge duplicates; keep one canonical record.

Practical workflow

  1. Set up the library before writing: folder structure or tags by project/topic.
  2. For each source read: save citation metadata, the file, and a 2-line note (claim + relevance).
  3. Write with cite keys; never type a formatted reference by hand.
  4. Before submission: run the integrity check — every key resolves, every entry is cited, no duplicates, style consistent.
  5. Verify AI-suggested citations individually: confirm the source exists and supports the claim.
  6. Archive the library with the manuscript version so the reference set is reproducible.
Source note template:
KEY:   smith2024attention
CLAIM: <what the paper shows, in your words>
USE:   <why you're citing it — which argument it supports>
READ:  <which sections you actually read>
LIMIT: <caveats relevant to your use>

Common pitfalls

  • Phantom citations: references that don't exist — the classic AI failure mode. Verify every single one.
  • Citation without reading: citing the abstract or a citation-of-a-citation. Read the part you rely on.
  • Hand-formatted references: typed bibliographies that break on every edit. Generate from keys.
  • No notes: a library of 500 PDFs with no memory of why each mattered. Annotate at capture time.
  • Style drift: mixed formats in one document. Set the style once; let the tool enforce it.
  • Uncited bibliography entries: sources in the list that nothing cites (or claims nothing supports). Audit both directions.

04 同仓库更多 Skills

TanNhatCMS
TanNhatCMS/jarvis

web-scrape

Turn web pages into LLM-ready markdown via the Firecrawl API (renders JS, anti-bot handling). Use when browser_open/fetch fails on JS-heavy pages, or when you need clean markdown/structured extraction from a site. Requires FIRECRAWL_API_KEY.

★ 0
TanNhatCMS
TanNhatCMS/jarvis

design-taste-frontend

tasteskill v2, the anti-slop frontend skill. Art direction for websites, distilled from a study of 247 curated gallery sites (One Page Love, Awwwards Sites of the Day). Deals researched directions (idea, composition, type, palette, photographs, section blueprints) from a library, enforces a strict no-slop house style with a real-browser inspector, and runs a look-and-fix loop on screenshots. Use this whenever you build, design, redesign, restyle, polish or review a landing page, marketing site, portfolio, product page, one-pager, hero or any public-facing web page, in any stack (HTML/CSS, React, Next.js, Vue, Svelte, Astro, Tailwind) — also when the user says a page looks generic, templated, "AI-made", sloppy, or asks to make it look designed, premium, award-level or less like every other site. Not for dashboards or app screens.

★ 0
TanNhatCMS
TanNhatCMS/jarvis

test-driven-development

Use when implementing any feature or bugfix, before writing implementation code

★ 0
TanNhatCMS
TanNhatCMS/jarvis

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes

★ 0
TanNhatCMS
TanNhatCMS/jarvis

requesting-code-review

Use when completing tasks, implementing major features, or before merging to verify work meets requirements

★ 0
TanNhatCMS
TanNhatCMS/jarvis

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).

★ 0