papers-query-knowledge-base

A search and summary tool for a local collection of analyzed research papers. It uses Markdown analysis notes and, when available, a generated index to find papers and compare their methods and evidence.

In plain words
What is it for?
It is for finding papers by title, topic, technique, venue, or year, then summarizing methods and evidence across the matching papers.
Why use it?
It removes the need to open and search research notes one by one. It also points back to the analysis metadata and linked papers so findings can be checked.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/ripemangobox/bite/papers-query-knowledge-base
Any agent
npx skills add RipeMangoBox/BITE --skill papers-query-knowledge-base
Clone the repo
git clone --depth 1 https://github.com/RipeMangoBox/BITE

Made for: Claude Code, Codex.

Per session 91 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,248 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00091 $0.01248
Opus 5 $0.00046 $0.00624
Sonnet 5 $0.00018 $0.00250
Haiku 4.5 $0.00009 $0.00125

Measured 2d ago against content hash dd33f0812a8c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

papers-query-knowledge-base scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/skills/papers-query-knowledge-base/SKILL.md · 76 lines

How it starts

The opening of the file, as written. The whole thing — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Paper Knowledge Base

Use this skill to query the local paper corpus across conversations and projects. The main evidence layer is obsidian-vault/analysis (analysis notes with TL;DR, semantic method/evidence sections, and PDF links). When it exists, obsidian-vault/index/index.jsonl is the fast filter layer for large vaults, while the Markdown pages in obsidian-vault/index/ remain useful for Obsidian jumps and backlink-friendly browsing.

Where the knowledge base lives: Under the current repository root that contains obsidian-vault/analysis/ and obsidian-vault/paperPDFs/. obsidian-vault/index/ is a generated companion layer created by papers-build-index.

Paths

  • Use repo-relative paths rooted at the folder that contains obsidian-vault/analysis/ and obsidian-vault/paperPDFs/.
  • When invoking from another workspace, replace these with the correct absolute repository path for your machine.

Where things live

Relative to the repository root:

  • Analysis notes: obsidian-vault/analysis/<Topic>/<Venue_Year>/<Year>_<Title>.md or flat venue folders such as obsidian-vault/analysis/ICLR_2026/<Title>.md; PDF path in frontmatter pdf_ref.
  • Generated index / navigation: obsidian-vault/index/index.jsonl, obsidian-vault/index/paper_index.md, _AllPapers.md, by_topic/, by_method/, by_dataset/, by_venue_year/.
  • Public placeholder: obsidian-vault/index/README.md is tracked as a stable explanation file and is not overwritten by index generation.

All paths use forward slashes.

How to use for research

  1. Find papers — Search obsidian-vault/analysis/ directly by title, task folder, tags, venue, year, core_operator, or primary_logic. If obsidian-vault/index/index.jsonl is present, use it first to narrow a large candidate set, then read the matching notes for evidence.

  2. Read an analysis — Open the matched analysis note directly. Each note has: Frontmatter (title, venue, year, tags, aliases, pdf_ref, core_operator, primary_logic, claims; no category, modalities, or frontier); Quick Links & TL;DR; method/problem/evidence sections such as 问题与动机, 整体框架, 核心模块与公式推导, 实验与分析, and 局限性与启发; and Local Reading (PDF). Older notes may still use legacy mixed headings, so match semantically rather than by one exact title string.

Read the full file on GitHub · 76 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 76 lines · 91 tokens per session scan A dd33f0812a8c

Subscribe to this mod's changes

papers-query-knowledge-base is a skill published in the GitHub repository RipeMangoBox/BITE (60 stars, last pushed 20d ago), licensed MIT. It adds 91 tokens to every session and 1,248 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.