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.
npx agentmods add instructions/o0000-code/paper-search-pro/claude-mdgit clone --depth 1 https://github.com/O0000-code/paper-search-proWrote 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.
[](https://agentmods.dev/instructions/o0000-code/paper-search-pro/claude-md)<a href="https://agentmods.dev/instructions/o0000-code/paper-search-pro/claude-md"><img src="https://agentmods.dev/badge/instructions/o0000-code/paper-search-pro/claude-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00713 | $0.00713 |
| Opus 5 | $0.00357 | $0.00357 |
| Sonnet 5 | $0.00143 | $0.00143 |
| Haiku 4.5 | $0.00071 | $0.00071 |
Grade A, and why
paper-search-pro CLAUDE.md 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 5d 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.
How it starts
The opening of the file, as written. The whole thing — 30 lines — stays where its author put it; the contents beside it link to each section on GitHub.
paper-search-pro — notes for sessions modifying this repo
Published Agent Skill (Apache 2.0). Edit the git repo here; the running copy
is separate at ~/.claude/skills/paper-search-pro/ — sync to it (rsync
scripts/ references/ assets/ + SKILL.md) for changes to take effect.
Invariants — do not break
- R-04 naming: OpenAlex / SJR / 中科院(CAS) journal metrics are partitions / open impact, never "Impact Factor". Only JCR
IF(YYYY)is a real Impact Factor. - R-19 human-path preservation: the human 14-STEP flow — especially STEP 6 (parallel classification SubAgents) and STEP 12 (HTML render) — must stay byte-identical when the agent/headless, source-switch, and journal-rank features are unused. Every new capability is opt-in additive; default = current behavior.
- Journal-rank data is never committed (Clarivate/SCImago/CAS copyright). It is runtime-fetched from public GitHub mirrors into
~/.paper-search-pro/ranks/.scripts/journal_rank.pyis the single canonical rank layer;scripts/sjr_helper.pyis superseded (kept for back-compat, not the wired source). - Zero new hard dependencies for v2.2 (GitHub-raw fetch uses
requests) — keep it that way: no playwright, no heavyweight libs.
Test
PYTHONPATH=. python3 -m pytest tests -q — ~264 offline pass. The ~12 failures in tests/test_openalex_helper.py are live OpenAlex rate-limit / key issues (environmental), not regressions.
Building the HTML report (bundle.html)
The report is TypeScript + Vite in assets/webartifacts_app/paper-report/. node_modules lives only
in the installed copy (~/.claude/skills/paper-search-pro/...) — the repo gitignores it — so build
there: rsync -a <repo>/…/src/ <inst>/…/src/ → cd <inst>/…/paper-report && pnpm exec vite build →
npx html-inline -i dist/index.html -o bundle.html → copy bundle.html back here (it's committed).
Use vite build, NOT pnpm build — the tsc -b gate is pre-existingly broken (React19/TS6 drift:
baseUrl deprecation + unused calendar.tsx/resizable.tsx type errors + an App.tsx union-props
strictness); none affect the runtime, and bundle.html has always been a vite+inline product.
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.
- 5d ago First seen · 30 lines · 713 tokens per session scan A 07780c030764
paper-search-pro CLAUDE.md is an instructions file published in the GitHub repository O0000-code/paper-search-pro (154 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 713 tokens to every session, about $0.0036 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.
Other instructions, from other repositories
paperbanana copilot-instructions.md
Instructions for llmsresearch/paperbanana, covering copilot instructions for paperbanana, build & test, install for development, run full test suite and run a single test file.
governed-agent-skills AGENTS.md
Instructions for Ezra144israel/governed-agent-skills, covering governed-agent-skills repository rules and repository rules.
governed-agent-skills CLAUDE.md
Instructions for Ezra144israel/governed-agent-skills: Read AGENTS.md. It governs this repository.
academic-tools-mcp CLAUDE.md
Instructions for hunter-heidenreich/academic-tools-mcp, covering claude.md, what this is, commands, code style & the format-on-edit hook and changelog & versioning.
sapcc-skill AGENTS.md
Instructions for eljoujat/sapcc-skill, covering agents.md, project overview, repository structure, key file: scripts/execute.js and execution flow.
fable-workflow-skill CLAUDE.md
Instructions for joey114132/fable-workflow-skill, covering fable-workflow-skill, layout, editing rules and adding a readme language.