SIE, or Superlinked Inference Engine, is a self-hosted inference server that runs the language, embedding, retrieval, and other models used by AI agents through one API. It is for teams deploying agent workloads in their own cloud and managing many models in a shared production cluster.
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 superlinked/sie --skill plugingit clone --depth 1 https://github.com/superlinked/sieWrote 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/superlinked/sie/plugin)<a href="https://agentmods.dev/skills/superlinked/sie/plugin"><img src="https://agentmods.dev/badge/skills/superlinked/sie/plugin/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/superlinked/sie/plugin"><img src="https://agentmods.dev/badge/skills/superlinked/sie/plugin.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 | $0.00125 | $0.02135 |
| Opus 5.5 | $0.00050 | $0.00854 |
| Sonnet 5.5 | $0.00025 | $0.00427 |
| Haiku 4.5 | $0.00013 | $0.00214 |
Grade A, and why
superlinked-docs 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 — 167 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Superlinked document offload
When the user gives you a document (PDF, DOCX, PPTX, XLSX, HTML, or a scan), do not open or attach the file into the conversation — that bills every page as text and a page-image. Convert it to markdown on the Superlinked cluster first, then work from the markdown.
docs → markdown
- Read the source file as raw bytes and base64-encode it. Do not view or attach it.
- Call the
docs_to_markdownMCP tool with:document_base64: the base64-encoded bytesfilename: the original filename (its extension hints the format)engine: leave asauto(default) — it runs the standard converter and automatically falls back to a vision OCR model for scanned/image-only pages, so scanned PDFs need no extra flag. Forcevl-ocrfor scanned or complex-layout PDFs you want OCR'd page-by-page, ordoclingto pin the standard converter.ocr: leavefalse. It only turns on the standard converter's built-in OCR underengine: docling; underautothe vision-OCR fallback handles scans, soocris ignored there.
- Write the returned
markdownto a file named after the source, then read and reason over that file — never the original. Where to write it depends on the surface:- Cowork (connected folder):
.superlinked/<original-name>.md(create.superlinked/if absent). - claude.ai / Claude desktop app (code-execution sandbox):
./<original-name>.mdin the working directory; offer to export it (Download, save to Google Drive, or add to a Project's knowledge to reuse it across chats). - Claude Code (local filesystem): alongside the source.
- Cowork (connected folder):
- From here on, read and reason over the saved
.md— never the original.
The tool response also includes a metadata object with two distinct figures: a live
per-call markdown_tokens_estimate (a rough ~4 chars/token count of this response — an
estimate, not a billed figure) and token_reduction, the reference benchmark
measurement of markdown vs direct document ingestion. Surface these if the user asks how
much was saved; the exact percentages and measurement run are in the metadata. Its
source is a stable, public-safe run label; issue is null because no public issue
tracks that committed measurement.
What ships with it
5 files 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.
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 First seen · 167 lines · 125 tokens per session scan A 27d3cdb066a9
superlinked-docs is a skill published in the GitHub repository superlinked/sie (3,402 stars, last pushed today), licensed Apache-2.0. It adds 125 tokens to every session and 2,135 once invoked, about $0.0005 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
meta-multi-format-export-pack
From one piece of source content, render four deliverables: .docx report, .pptx slides, .xlsx data, and an HTML/PDF public version.
pdf-toolkit
Structured .pdf operations: extract text/tables, merge pages from multiple PDFs, split a PDF by page ranges, fill PDF form fields, and generate fresh PDFs from JSON. Trigger when the user wants programmatic PDF work without natural-language rewriting — examples: pull tables from a report, combine three PDFs, extract…
xlsx
Read, edit, or create Microsoft Excel .xlsx workbooks. Trigger this skill whenever the user mentions a spreadsheet, .xlsx file, workbook, sheet, formula, pivot table, or asks to extract tabular data, modify a sheet, or build a workbook from rows. Three execution paths: structured inspection, in-place cell edits, and…
meta-pdf-intelligence
Use this meta-skill instead of answering directly when the user needs PDF analysis, pasted PDF excerpt analysis, digesting, comparison, or question answering that benefits from multi-skill orchestration across PDF extraction, summarization, cross-document synthesis, traceable evidence indexing, and memory capture.
smart-data-collection
A workflow for extracting structured information from images and documents such as PDFs, Word files, and spreadsheets, then storing it in a database.
meta-research-to-slide-deck
Use this meta-skill instead of answering directly when the user needs a researched presentation, leadership briefing, competitive analysis deck, or source-backed slide outline that benefits from multi-skill orchestration across search, source curation, synthesis, slides, and document export.