- Home
- Skills
- Agents & Orchestration
- Axiom Research
Axiom Research
A modular LLM-assisted research framework connecting scientific evidence, numerical modelling and experimental feedback through reusable Skills for traceable materials and process research.
Free
Works with the AI Agents you already use
Axiom Research
Example session with this skill installed
Use the bundled sparse-data VPP–DIW example to assess calibration readiness and prepare a three-trial plan within the declared bounds. Label the output as a synthetic software demonstration. If validated predictions are unavailable, explain what measurements are needed.
- Read your context and instructions
- Compiled the axiom research
Synthetic software demonstration
English summary of the pre-existing bundled report in reports/sparse-v1/report.md. The source is labelled synthetic_demo and deterministic_demo. The values below are artificial software-test fixtures.
The recorded workflow reached awaiting_external_experiment. Input inspection, evidence checking, calibration assessment and planning were completed. No external feedback was attached, imported or compared, and no model update was prepared.
Recorded candidate trials
| Trial | Purpose | Speed (mm/s) | Delay (s) | Frozen prediction (µm) |
|---|---|---|---|---|
| T03 | Baseline replication | 5.0 | 3.0 | None |
| T01 | Coverage or calibration | 15.0 | 2.0 | None |
| T02 | Coverage or calibration | 13.0 | 5.0 | None |
The report contains no numerical predictions for these trials. The settings belong to the synthetic fixture and require researcher review before any physical experiment. The software did not approve or execute the trials.
Next step and traceability
The workflow calls finish_report and waits for the operator to attach new reported measurements. Requests, responses, call origins, hashes and errors are recorded in events.jsonl, with structured results retained in calls/.
This bundled deterministic example demonstrates software behaviour. Physical experiments and live LLM execution are outside this recorded example.
Connects securely to your tools. The creator never sees your data.
About this skill
Axiom Research is a modular LLM-assisted research framework that connects scientific evidence, numerical modelling and experimental feedback through reusable Skills. It helps researchers turn research questions, available data and practical constraints into traceable analyses and proposals for materials-design and process-improvement studies.
A file- and terminal-capable LLM host organises evidence and selects reviewed Skills at each technical stage, while numerical tools perform the calculations. Depending on the chosen module, the workflow supports evidence review, model calibration, computational experiments, candidate assessment and discrepancy analysis. Researchers can inspect the inputs, assumptions and results before deciding which physical experiments to carry out.
Current application modules
-
VPP–DIW track width and diffusion: separate empirical geometric-width and concentration-sigma routes, with calibration, grouped validation, bounded experiment proposals, measured-feedback comparison against frozen predictions, and versioned model updates subject to fresh confirmation.
-
Hierarchical polymer foams: an independent module for lightweight two-level modelling, calibration, predeclared model refinements and limited sensitivity analysis within supported domains.
The package includes reusable Skills, numerical scripts, research templates, synthetic demonstration fixtures and documentation. Its modular interfaces provide a foundation for extending the workflow with suitable data, models and reviewed adapters.
Scope and requirements
The implemented models have documented material, observable and process limits. The foam module has its own entry point and does not inherit the VPP–DIW multi-round workflow. Bundled numerical demonstrations use synthetic data. Physical experiments and research decisions remain under researcher control.
Requires Python 3.11+, the dependencies specified for the selected module, and an LLM host with file and terminal access.
Source and documentation: https://github.com/yuerway983-create/axiom-research-skill Project-owned files are available under the MIT licence, with third-party attribution retained in the package.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Models apply within their documented material, observable and process scope and require suitable data. Bundled fixtures are synthetic software demonstrations. Suggested physical experiments require researcher review and external execution. The foam module supports independent modelling and calibration; automatic experiment planning and an integrated multi-round loop are outside its current scope. Local tests do not certify live API integration.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 11 days ago
- Free with an Agensi account
- Software
Python 3.11+Terminal accessPython dependencies
Node.js
Git
Install the dependencies from requirements-foam.txt or requirements-agent.txt.
Permissions
Local analysis reads project files and researcher-supplied data, runs Python tools, and writes reports and versioned outputs. Browser or network access is used only for authorised literature/source retrieval or the optional API driver. Environment variables and an OpenAI account are required only when using the optional OpenAI Responses driver.