MMADS commited on
Commit
6b7cbd6
·
1 Parent(s): dbebdd4

final cosmetic changes

Browse files
Files changed (1) hide show
  1. app.py +14 -17
app.py CHANGED
@@ -1,4 +1,4 @@
1
- #!/usr/bin/env python
2
  """CVE Dashboard - Real-time vulnerability monitoring with NVD API and LLM-powered audience customization."""
3
 
4
  import os
@@ -389,7 +389,7 @@ def generate_tailored_summary(cve_description: str, audience: str, hf_token: Opt
389
  profile = AUDIENCE_PROFILES[audience]
390
 
391
  # Gemma uses a specific chat template format.
392
- # We combine the system and user prompts into a single user turn.
393
  full_prompt = f"""You are an expert cybersecurity analyst.
394
 
395
  Rewrite this CVE description for a {audience}.
@@ -587,7 +587,7 @@ def create_interface():
587
  """
588
  ### About this Dashboard
589
 
590
- This dashboard provides real-time monitoring of Common Vulnerabilities and Exposures (CVEs)
591
  from the NIST National Vulnerability Database with AI-powered audience customization.
592
 
593
  **Features:**
@@ -595,6 +595,12 @@ def create_interface():
595
  - Filter by severity levels
596
  - Visualize CVE distributions and trends
597
  - AI-powered audience-specific summaries using the google/gemma-2-2b-it model.
 
 
 
 
 
 
598
 
599
  **Supported Audiences:**
600
  - **Cybersecurity Professional:** Focus on threats, attack vectors, and mitigation
@@ -610,21 +616,12 @@ def create_interface():
610
 
611
  **Disclaimer:** Generated content may be inaccurate or false.
612
 
613
- The free community tier of the Hugging Face Inference API powers this app's AI features. Since computing resources are shared, anticipate some delay on your initial request as the model loads. Later requests usually process more quickly.
614
-
615
- **Performance Optimizations:**
616
- - Shorter timeouts for faster failure detection
617
- - Optimized prompts for quicker responses
618
 
619
- **Rate Limits:**
620
- - NVD API: 5 requests/30s (without key), 50 requests/30s (with key)
621
- - HuggingFace API: Varies by plan, includes cold start delays
622
-
623
- **Severity Levels:**
624
- - **CRITICAL** (9.0-10.0): Complete system compromise possible
625
- - **HIGH** (7.0-8.9): Significant impact, immediate patching recommended
626
- - **MEDIUM** (4.0-6.9): Moderate impact, plan for updates
627
- - **LOW** (0.1-3.9): Minor impact, update in regular cycle
628
  """
629
  )
630
 
 
1
+
2
  """CVE Dashboard - Real-time vulnerability monitoring with NVD API and LLM-powered audience customization."""
3
 
4
  import os
 
389
  profile = AUDIENCE_PROFILES[audience]
390
 
391
  # Gemma uses a specific chat template format.
392
+ # Combine the system and user prompts into a single user turn.
393
  full_prompt = f"""You are an expert cybersecurity analyst.
394
 
395
  Rewrite this CVE description for a {audience}.
 
587
  """
588
  ### About this Dashboard
589
 
590
+ This dashboard provides real-time monitoring of [Common Vulnerabilities and Exposures (CVEs)](https://en.wikipedia.org/wiki/Common_Vulnerabilities_and_Exposures)
591
  from the NIST National Vulnerability Database with AI-powered audience customization.
592
 
593
  **Features:**
 
595
  - Filter by severity levels
596
  - Visualize CVE distributions and trends
597
  - AI-powered audience-specific summaries using the google/gemma-2-2b-it model.
598
+
599
+ **Severity Levels:**
600
+ - **CRITICAL** (9.0-10.0): Complete system compromise possible
601
+ - **HIGH** (7.0-8.9): Significant impact, immediate patching recommended
602
+ - **MEDIUM** (4.0-6.9): Moderate impact, plan for updates
603
+ - **LOW** (0.1-3.9): Minor impact, update in regular cycle
604
 
605
  **Supported Audiences:**
606
  - **Cybersecurity Professional:** Focus on threats, attack vectors, and mitigation
 
616
 
617
  **Disclaimer:** Generated content may be inaccurate or false.
618
 
619
+ The free community tier of the Hugging Face Inference API powers this app's AI features.
620
+ Since computing resources are shared, anticipate some delay on your initial request as the model loads. Later requests usually process more quickly.
 
 
 
621
 
622
+ **Note:** If you encounter rate limiting or timeouts, please try again after a short wait.
623
+
624
+ **Developed by** [M. Murat Ardag](https://mmuratardag.github.io/).
 
 
 
 
 
 
625
  """
626
  )
627