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 skills add feniix/pi-extensions --skill research-paper-searchgit clone --depth 1 https://github.com/feniix/pi-extensionsWrote 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/skills/feniix/pi-extensions/research-paper-search)<a href="https://agentmods.dev/skills/feniix/pi-extensions/research-paper-search"><img src="https://agentmods.dev/badge/skills/feniix/pi-extensions/research-paper-search/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.
<a href="https://agentmods.dev/skills/feniix/pi-extensions/research-paper-search"><img src="https://agentmods.dev/badge/skills/feniix/pi-extensions/research-paper-search.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 5.1 | $0.00015 | $0.00592 |
| Opus 5.5 | $0.00006 | $0.00237 |
| Sonnet 5.5 | $0.00003 | $0.00118 |
| Haiku 4.5 | $0.00002 | $0.00059 |
Grade A, and why
research-paper-search 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 27d 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 — 45 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Research Paper Search (Exa)
Use this skill for discovering academic papers, arXiv/preprint work, methods comparisons, literature scans, and evidence-weighted technical summaries.
Tool Selection
| Intent | Primary Tool | Notes |
|---|---|---|
| Paper discovery | web_search_advanced_exa with category: "publication" |
Use when available; domain filters such as arxiv.org can improve signal. |
| Broad discovery when advanced search is unavailable | web_search_exa |
Include topic, method names, and paper/source terms. |
| Evidence-weighted synthesis across papers | web_research_exa |
Use when enabled; specify evidence quality and caveat requirements. |
| One-off definition or quick answer | web_answer_exa |
Keep it concise and citation-focused. |
| Read abstracts or paper pages | web_fetch_exa |
Fetch selected paper URLs when snippets are insufficient. |
| Find related papers from one seed paper | web_find_similar_exa |
Use when a seed URL is clearly on-topic. |
If opt-in tools are unavailable, use web_search_exa plus selective web_fetch_exa and produce a smaller sourced summary.
Recommended Settings
- Paper discovery
{ "query": "LLM fine-tuning methods instruction tuning preference optimization", "category": "research paper", "includeDomains": ["arxiv.org"], "numResults": 10 }
- Related-paper expansion
{ "url": "https://arxiv.org/abs/2305.18290", "excludeSourceDomain": false, "numResults": 8 }
- Evidence-weighted synthesis
{ "query": "Summarize methodological differences between instruction tuning, RLHF, DPO, and newer preference optimization methods for large language models.", "systemPrompt": "Prioritize peer-reviewed papers, arXiv papers with clear methods sections, and papers with empirical comparisons. Separate findings from speculation and call out methodology limits.", "outputSchema": { "type": "text" } }
Query Writing
- Include method names, task/domain, and publication terms:
paper,arXiv,benchmark,survey,method. - Use date filters for fast-moving areas when the user asks for recent work.
- Prefer direct paper sources over secondary summaries unless the user wants a high-level introduction.
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.
- 27d ago Changed 45206dd524a7
- 1mo ago First seen · 45 lines · 15 tokens per session scan A dc1e311133b6
research-paper-search is a skill published in the GitHub repository feniix/pi-extensions (25 stars, last pushed 3d ago), licensed MIT. It adds 15 tokens to every session and 592 once invoked, about $0.0001 per session on Opus 5.5. 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 skills, from other repositories
using-pi-subagents
Operate pi-subagents jobs safely, including direct-work decisions, least-privilege tools and attachments, thinking levels, delegation, bidirectional messaging, parallel starts, timeouts, waiting, cancellation, result handling, verification, and writer isolation.
configure-pi-starship
Configure @narumitw/pi-starship and answer questions about its pi-starship.toml settings. Use when the user asks to create, edit, repair, migrate, or understand pi-starship.toml, or asks which pi-starship module, variable, option, style, palette, preset, or runtime behavior to use. Do not use for generic TOML, shell…
codebase-memory
Use the codebase knowledge graph for structural code queries. Triggers on: explore the codebase, understand the architecture, what functions exist, show me the structure, who calls this function, what does X call, trace the call chain, find callers of, show dependencies, impact analysis, dead code, unused functions…
typesafe-ai
Build AI-powered software with TypeSafe and Jev, or use pi-typesafe's typesafequestion tool for typed semantic judgments and probabilities. Use when a feature needs programmable common sense, when an LLM prompt-and-parse step could become a structured decision, or when designing, calling, evaluating, or…
configure-pi-sync
Explain @narumitw/pi-sync settings and guide safe setup through its existing manager. Use when the user asks how to configure pi-sync.json, choose Git, WebDAV, R2, or S3 storage, select synced content, understand automatic controls, or recover unsupported settings. This skill provides guidance only, not settings edits…
pi-observability-local-dev
Use when developing or reviewing packages/pi-observability dashboard/server changes locally, especially rebuilding dist, restarting the 43190 server, validating UI behavior, or keeping the local dashboard in sync with source edits.