Back to skill

Security audit

Bohrium LKM (Large Knowledge Model)

Security checks across malware telemetry and agentic risk

Overview

This is a purpose-aligned Bohrium API helper, but users should treat submitted research queries, claims, paper IDs, and OCR requests as data sent to a third-party service.

Install only if you are comfortable using Bohrium as a third-party processor for your scientific searches, claims, paper identifiers, and OCR tasks. Avoid submitting confidential manuscripts, proprietary hypotheses, regulated data, or restricted/licensed documents unless you have approval and understand Bohrium's data handling terms.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (7)

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill is explicitly designed to send scientific queries, claims, and paper identifiers to a third-party API, but it does not warn users that their inputs will leave the local environment and be transmitted to open.bohrium.com. This creates a real data-handling risk because users may submit unpublished research ideas, proprietary claims, or sensitive document identifiers without informed consent.

External Transmission

Medium
Category
Data Exfiltration
Content
## 1. Knowledge graph search — `/lkm/search`

```python
r = requests.post(f"{BASE}/search", headers=H, json={
    "query": "effect of temperature on lithium ion battery degradation",
    "limit": 10
})
Confidence
95% confidence
Finding
requests.post(f"{BASE}/search", headers=H, json=

External Transmission

Medium
Category
Data Exfiltration
Content
Submit a scientific claim, get back evidence that supports or refutes it (with source papers and relevance scores).

```python
r = requests.post(f"{BASE}/claims/match", headers=H, json={
    "text": "Graphene oxide improves the mechanical strength of concrete",
    "limit": 5
})
Confidence
97% confidence
Finding
requests.post(f"{BASE}/claims/match", headers=H, json=

External Transmission

Medium
Category
Data Exfiltration
Content
Batch OCR extraction from papers.

```python
r = requests.post(f"{BASE}/papers/ocr/batch", headers=H, json={
    "paper_ids": ["doi:10.1038/s41586-021-03819-2", "doi:10.1126/science.abf3041"]
})
data = r.json()
Confidence
96% confidence
Finding
requests.post(f"{BASE}/papers/ocr/batch", headers=H, json=

External Transmission

Medium
Category
Data Exfiltration
Content
---

## curl examples

```bash
AK="YOUR_ACCESS_KEY"
Confidence
88% confidence
Finding
curl examples ```bash AK="YOUR_ACCESS_KEY" # Knowledge graph search curl -s -X POST "https://open.bohrium.com/openapi/v1/lkm/search" \ -H "accessKey: $AK" -H "Content-Type: application/json" \ -

External Transmission

Medium
Category
Data Exfiltration
Content
-d '{"query":"lithium battery degradation mechanism","limit":10}' | jq .

# Claim matching
curl -s -X POST "https://open.bohrium.com/openapi/v1/lkm/claims/match" \
  -H "accessKey: $AK" -H "Content-Type: application/json" \
  -d '{"text":"MoS2 is a promising catalyst for hydrogen evolution","limit":5}' | jq .
Confidence
94% confidence
Finding
curl -s -X POST "https://open.bohrium.com/openapi/v1/lkm/claims/match" \ -H "accessKey: $AK" -H "Content-Type: application/json" \ -d

External Transmission

Medium
Category
Data Exfiltration
Content
-d '{"ids":["gcn_b2bf079b541a4fa0","gcn_5cecd02c3d8a4e61"]}' | jq .

# Batch OCR
curl -s -X POST "https://open.bohrium.com/openapi/v1/lkm/papers/ocr/batch" \
  -H "accessKey: $AK" -H "Content-Type: application/json" \
  -d '{"paper_ids":["doi:10.1038/s41586-021-03819-2"]}' | jq .
```
Confidence
95% confidence
Finding
curl -s -X POST "https://open.bohrium.com/openapi/v1/lkm/papers/ocr/batch" \ -H "accessKey: $AK" -H "Content-Type: application/json" \ -d

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.