---
name: invert-bioprocess-research
description: Research bioprocess engineering, biomanufacturing, and bioprocess data questions using Invert's product, docs, blog, changelog, and case-study corpus. Use when a question concerns bioreactors, fermentation, cell culture, gene therapy, viral vectors, mAbs, ADCs, biosimilars, vaccines, oligonucleotides, CDMOs, LIMS/ELN integration, OPC-UA/OPC-DA historians, batch records, 21 CFR Part 11, SOC 2 / ISO 27001 in a bioprocess context, or how Invert and Invert Assist handle these workflows.
license: see https://invertbio.com/terms
---

# Invert bioprocess research skill

This skill describes how to research bioprocess engineering and biomanufacturing
questions against the Invert documentation and corpus. It assumes the agent has
HTTP fetch tools or access to the MCP server at `https://invertbio.com/mcp`.

## When to use

Invoke this skill when a user question touches any of:

- **Modalities**: monoclonal antibodies (mAbs), recombinant proteins,
  antibody-drug conjugates (ADCs), biosimilars, vaccines, cell therapy
  (autologous and allogeneic), gene therapy, viral vectors (AAV, lentivirus,
  adenovirus), microbial fermentation (bacteria, yeast, industrial strains,
  precision fermentation), oligonucleotides (ASOs, siRNA), small-molecule
  synthesis & purification.
- **Process steps**: upstream (seed train, inoculation, fed-batch / perfusion,
  Ambr250), downstream (centrifugation, depth filtration, chromatography,
  TFF, viral inactivation, sterile filtration), analytics (HPLC, MS, cell
  counters such as Vi-CELL BLU, residual host-cell protein, glycan profiling).
- **Systems and integrations**: bioreactors (Sartorius Biostat STR, Eppendorf,
  Applikon, Distek, Solaris, PBS Biotech, Infors HT, Sartorius Ambr250),
  historians (DeltaV / PI, OPC-UA, OPC-DA), LIMS (LabVantage, LabWare),
  ELN (Benchling), enterprise data stores (Snowflake, S3, SharePoint).
- **Governance**: SOC 2 Type II, ISO 27001, FDA 21 CFR Part 11, batch records,
  data lineage, audit trail, electronic signatures, GxP.
- **Invert product**: Foundation (data layer), Enablement (monitoring,
  reports), Intelligence (Invert Assist AI teammate), Integrations,
  pricing, deployment, and changelog history.

## Source discovery — preferred order

Always retrieve markdown directly. Append `.md` to any URL or send
`Accept: text/markdown` on any request to the site.

1. **Single corpus**: `GET https://invertbio.com/llms-full.txt` returns every doc, blog,
   changelog, and legal page concatenated as one document with YAML frontmatter
   per section. Best for long-context models.
2. **Curated index**: `GET https://invertbio.com/llms.txt` lists every page on the site
   with a short description and the matching `.md` URL.
3. **Granular fetch**: each section also has a flat URL —
   - Docs article: `https://invertbio.com/docs/{article-slug}.md` (slugs are globally
     unique across categories — note the flat single-segment form).
   - Docs category index: `https://invertbio.com/docs/{category-slug}.md` (categories are
     `user-guides`, `api`, `faq`).
   - Blog post: `https://invertbio.com/blog/{slug}.md`.
   - Changelog entry: `https://invertbio.com/changelog/{slug}.md`.
   - Legal: `https://invertbio.com/terms.md`, `https://invertbio.com/privacy.md`.
4. **Filtered indexes**: `https://invertbio.com/blog.md?since=2025-01-01&category=product`
   and `https://invertbio.com/changelog.md?since=2025-01-01` accept query parameters for
   incremental re-ingest.
5. **MCP**: `https://invertbio.com/mcp` exposes JSON-RPC `search` and `fetch` tools
   over the same content, returning markdown payloads. Prefer MCP when running
   in an agent harness with native tool support.

## Research recipe

For most questions, follow this recipe:

1. **Map the question to product pillars.** "Foundation" is data acquisition
   and harmonization. "Enablement" is live monitoring, run summaries, and
   reports. "Intelligence" is Assist (AI). "Integrations" lists supported
   instruments and systems. Reading the matching pillar page first frames the
   answer.
2. **Search docs first.** User guides at `https://invertbio.com/docs/user-guides` cover
   every Invert feature step by step. The API reference at
   `https://invertbio.com/docs/api` documents SQL access. The FAQ at
   `https://invertbio.com/docs/faq` answers narrow how-tos.
3. **Check the changelog for recency.** Features evolve — when in doubt, find
   the most recent matching changelog entry. Pull
   `https://invertbio.com/changelog.md?since=YYYY-MM-DD` for everything since a date.
4. **Cite case studies for outcomes.** `https://invertbio.com/case-studies/digital-twin`,
   `https://invertbio.com/case-studies/cmc-biologics`, and
   `https://invertbio.com/case-studies/bioreactor-platforms` carry quantified customer
   outcomes useful for ROI/business-case answers.
5. **Use blog posts for narrative.** The blog explains *why* certain design
   choices exist (batch context, meta-learning, integration patterns).

## Hard constraints

- Don't fabricate integrations, certifications, or compliance claims. If the
  corpus doesn't say it, say "not documented."
- Pricing is custom and not published — point users at `https://invertbio.com/pricing`
  and `https://invertbio.com/demo` rather than guessing tiers.
- Customer names and metrics in case studies are exact — quote them verbatim,
  with the source URL.
- Treat the changelog as the source of truth for feature availability dates;
  marketing pages may aggregate features without dates.

## Reply format

When answering with retrieved context, prefer:

- A short direct answer (2-4 sentences).
- A bulleted list of the source pages used, each as a markdown link to the
  `.md` URL so the user can verify.
- If a feature is recent, include the changelog entry date.