# Molclaw Boltz2 Affinity

> Predict binding affinity between target protein sequence and small molecule SMILES using Boltz-2.

- **Type:** Skill
- **Install:** `agentstack add skill-internscience-molclaw-molclaw-boltz2-affinity`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [InternScience](https://agentstack.voostack.com/s/internscience)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [InternScience](https://github.com/InternScience)
- **Source:** https://github.com/InternScience/MolClaw/tree/main/skills/L1_tools/molclaw-boltz2-affinity

## Install

```sh
agentstack add skill-internscience-molclaw-molclaw-boltz2-affinity
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Boltz-2 Protein-Ligand Binding

Note: 
- Local files are not directly accessible by the server. Please upload them to the server using `molclaw-file-transfer` before execution. 
- For PDB file inputs, it is recommended to preprocess them using `molclaw-pdbfixer` before execution.
- Please refer to skill `molclaw-scp-server` to complete tool invocation.

step 1. Use skill **molclaw-protein-sequence-retrieve** to get the target protein sequence information. If the target protein sequence has been provided, skip this step.

step 2. Finally use tool *pred_binding_affinity_boltz2* to predict the binding affinity.  

Tool description:

```tex
Use Boltz to predict binding affinity between protein (receptor) and small molecule (ligand).
Args:
    protein (List[dict]): Protein chains, each element contains 'chain' and 'sequence' (e.g., [{{'chain': 'A', 'sequence': 'MGNAAAAKKGSEQASQRRSSLEQP*'}}])
    smiles (str): Input SMILES string (e.g., "N[C@@H](Cc1ccc(O)cc1)C(=O)O")
Return:
    status (str): success/error
    msg (str): message
    affinity_probability_binary (float): Represents the predicted probability (ranging from 0 to 1) that a ligand is a binder, making it ideal for distinguishing active compounds from decoys during the hit-discovery stage. A value below 0.5 indicates uncertain or weak binding.
    affinity_pred_value (float): Estimates the specific binding affinity as log10(IC50) in μM to quantify how small molecular modifications affect potency, serving as a key metric for ligand optimization phases like hit-to-lead and lead-optimization.
    complex_cif_file (str): Structure file of the protein–molecule complex
```

Tool usage:

```python
response = await client.session.call_tool(
    "pred_binding_affinity_boltz2",
    arguments={
        "protein": protein_chains,
        "smiles": smiles
    }
)
result = client.parse_result(response)
affinity_probability_binary = result["affinity_probability_binary"]
affinity_pred_value = result["affinity_pred_value"]
```

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [InternScience](https://github.com/InternScience)
- **Source:** [InternScience/MolClaw](https://github.com/InternScience/MolClaw)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-internscience-molclaw-molclaw-boltz2-affinity
- Seller: https://agentstack.voostack.com/s/internscience
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.