obsidian-llm-wiki: Command for Claude Code

.claude/commands/paper-ask.md

paper-ask is a command for Claude Code from mickeytony0215-png/obsidian-llm-wiki. It costs 33 tokens per session (700 once invoked), scanned B, original, MIT.

A command for asking questions across the full text of a library of research-paper PDFs. It retrieves supporting passages and cites the paper and page for each claim, but requires a separate search system such as RAGFlow.

In plain words
What is it for?
Use it to find papers about a topic, compare methods and baselines, locate reported measurements, and support related-work writing.
Why use it?
It helps avoid relying on filenames or memory when comparing research or checking details. Page citations make the answers easier to verify.

Command for Claude Code

Written for Claude Code: installed under .claude/.

This is mickeytony0215-png/obsidian-llm-wiki's own configuration. It tells Claude Code how to work on obsidian-llm-wiki itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything obsidian-llm-wiki configures →

Reuse

Borrowing it

Nothing to install: this file belongs to mickeytony0215-png/obsidian-llm-wiki. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/mickeytony0215-png/obsidian-llm-wiki/main/.claude/commands/paper-ask.md
Clone the repo
git clone --depth 1 https://github.com/mickeytony0215-png/obsidian-llm-wiki

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for paper-ask

README.md
[![agentmods](https://agentmods.dev/badge/commands/mickeytony0215-png/obsidian-llm-wiki/paper-ask/github.svg)](https://agentmods.dev/commands/mickeytony0215-png/obsidian-llm-wiki/paper-ask)
Your own site
<a href="https://agentmods.dev/commands/mickeytony0215-png/obsidian-llm-wiki/paper-ask"><img src="https://agentmods.dev/badge/commands/mickeytony0215-png/obsidian-llm-wiki/paper-ask/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for paper-ask

Your own site · 80×15
<a href="https://agentmods.dev/commands/mickeytony0215-png/obsidian-llm-wiki/paper-ask"><img src="https://agentmods.dev/badge/commands/mickeytony0215-png/obsidian-llm-wiki/paper-ask.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 33 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 700 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00033 $0.00700
Opus 5 $0.00016 $0.00350
Sonnet 5 $0.00007 $0.00140
Haiku 4.5 $0.00003 $0.00070

Measured 12d ago against content hash e65db94b4883, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade B, and why

paper-ask scanned grade B with 1 finding 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 12d 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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

> **This repo ships the command spec only.** The retrieval backend is intentionally not bundled — see [RAGFlow](https://github.com/infiniflow/ragflow) for the reference implementation this spec was extracted from. Fork u
.claude/commands/paper-ask.md · 32 lines

How it starts

The opening of the file, as written. The whole thing — 32 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/paper-ask — ask a question of the whole paper library

This repo ships the command spec only. The retrieval backend is intentionally not bundled — see RAGFlow for the reference implementation this spec was extracted from. Fork users either stand up their own self-hosted retrieval stack (RAGFlow or an equivalent chunked-PDF index) or swap this command for a simpler pdftotext + grep flow. .claude/scripts/ragflow_*.py (ingest / chunk-dump helpers) are local infrastructure and are not shipped, consistent with this repo's spec-only stance on scripts (see /search-vault).

Grounded Q&A over the full-text index of every PDF under raw/papers/. Use it for:

  • Related-work drafting: "Which papers use method X for problem Y? How do their designs differ?"
  • Finding a specific number: "Which papers report a latency measurement for Z?"
  • Cross-checking: "Do these three papers' baselines overlap?"
  • Recall by content, not by filename: "Which paper discussed the multi-objective optimization for RSU placement?"

Procedure

  1. Retrieve. Query the corpus-wide retrieval tool (e.g. a RAGFlow MCP retrieval call, or the equivalent HTTP/CLI endpoint of whatever local stack is standing in for it). If one phrasing doesn't surface enough evidence, retry with 2–3 different phrasings (synonyms, alternate terminology, another language) before concluding the corpus has nothing.
    • Retrieval backend not running: tell the user how to start it, then stop — do not silently fall back to un-grounded answers.
  2. Answer, grounded only in retrieved evidence:
    • Answer strictly from what the retrieved chunks contain. If they don't cover the question, say so explicitly ("not found in the corpus") — never fill the gap from the model's own training knowledge without clearly labeling it as "supplementary, outside the corpus."
    • Every claim carries its source: paper filename/slug + page number, taken from the retrieval evidence's own page metadata.
    • Quote numeric data as it appears in the evidence — do not round or paraphrase figures.
  3. Scope reminder. The corpus only covers PDFs that have actually been ingested. If the user asks about a paper that clearly isn't indexed (no hits, and no matching file under raw/papers/), suggest /read-paper or the ingest script to add it first.

Read the full file on GitHub · 32 lines

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. 12d ago First seen · 32 lines · 33 tokens per session scan B e65db94b4883

Subscribe to this mod's changes

paper-ask is a command published in the GitHub repository mickeytony0215-png/obsidian-llm-wiki (2 stars, last pushed 2mo ago), licensed MIT. It adds 33 tokens to every session and 700 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.