synthetic-sciences/openscience is an AI workbench that carries out scientific research by reading papers, forming hypotheses, writing and running code, conducting experiments, analyzing results, and preparing reports. Researchers use it for work in machine learning, biology, physics, and chemistry with remote or local models. Catalogue add-ons extend its scientific workflows through skills and instructions.
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 synthetic-sciences/openscience --skill sourcesgit clone --depth 1 https://github.com/synthetic-sciences/openscienceWrote 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/synthetic-sciences/openscience/sources)<a href="https://agentmods.dev/skills/synthetic-sciences/openscience/sources"><img src="https://agentmods.dev/badge/skills/synthetic-sciences/openscience/sources/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/synthetic-sciences/openscience/sources"><img src="https://agentmods.dev/badge/skills/synthetic-sciences/openscience/sources.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00097 | $0.01068 |
| Opus 5.5 | $0.00039 | $0.00427 |
| Sonnet 5.5 | $0.00019 | $0.00214 |
| Haiku 4.5 | $0.00010 | $0.00107 |
Grade A, and why
sources 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 1mo 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 — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Sources audit
A sources audit answers one question for every claim: where exactly does this come from, and does that place say this? It is the check that catches confident paraphrase drifting from what the source said, numbers copied with the wrong units, and results attributed to the wrong paper. The deliverable is a table, not a reassurance.
Rules
- Atomize. Split the text into claims small enough to be true or false on their own. "X improves accuracy by 4% on Y and generalizes to Z" is two claims.
- Locate, then read. For each claim, find the exact supporting passage: a sentence, a table cell, a figure panel, a dataset field, a tracked run's metric. Read it in its context; a sentence taken from a limitations paragraph often says the opposite of the claim built on it.
- Grade against the wording. Supported: the source says this, in this scope. Partially supported: the source says something narrower, weaker or conditional. Unsupported: no source found, or the source does not address it. Contradicted: the source says otherwise. Recall is never a source.
- Numbers are exact. Value, units, precision, uncertainty and conditions must match. A rounded or unit-converted number is partially supported unless the text says it converted.
- Attribution is part of the claim. A correct fact attributed to the wrong paper is contradicted, not supported.
- Report, do not repair. The audit lists what it found; changing the text is a separate step the user asks for (paper-writing).
Workflow
- Collect the claims (numbered) and the sources the text cites or relies on.
- For each claim, locate and read the supporting passage; fetch sources not at hand.
- Grade each claim; record the passage and location.
- Summarize: counts by grade, the claims that matter most, what to fix first.
Locating. Cited papers: fetch by DOI or arXiv id (requests in
skills/core/citations/references/apis.md) and read the relevant section. Project
results: open the output file or query the tracked run (experiments compare,
experiments series). Datasets: the field and record. Web sources: the page as of today,
with its date.
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.
- 1mo ago Changed · +6 lines 7524c8bcd1c8
- 1mo ago First seen · 82 lines · 97 tokens per session scan A 130b09a89114
sources is a skill published in the GitHub repository synthetic-sciences/openscience (4,025 stars, last pushed today), licensed Apache-2.0. It adds 97 tokens to every session and 1,068 once invoked, about $0.0004 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 2026-09-14.
Other skills, from other repositories
knowledge-extension-ecosystem-lsp-code-intelligence
A codebase guide to LSP, the standard protocol behind editor features such as symbol queries, file diagnostics, and language-aware code information.
github-explorer
Explore and summarize popular GitHub projects. Use when the user wants to discover trending GitHub repositories, analyze project details, or get summaries of popular projects.
google-workspace-setup
One-time setup for gws: install, OAuth, scopes, auto-approve.
google-sheets
Google Sheets via gws: read/write cells, append rows, structured batch edits.
attachments
Move bytes between Gini upload space, external URLs, and workspace files. Used by every attachment / file-upload / file-download flow regardless of the target system (Linear, GitHub, S3, Notion, etc.).
google-docs
Google Docs via gws: read, append text, structured batch edits.