Requirements
This page lists the baseline requirements to run Qualys AI Triage Pack successfully.
System requirements
Python: 3.11 or newer
Operating system: Windows, macOS, or Linux (any OS that supports Python 3.11+)
Disk space:
Enough space for input CSVs and generated outputs (
output/<csv_name>/...)Additional space if you enable ZIP attachments/uploads
Network access:
Required if you use the LLM (access to your configured OpenAI endpoint)
Required if you enable any external integrations (Slack, SendGrid, OpenSearch, Splunk, syslog server, Postgres)
Input requirements
One or more Qualys CSV exports (the pipeline ingests CSV files from the configured input directory).
Core runtime dependencies
Installed automatically when you install the package (see pyproject.toml / requirements.txt):
Configuration & CLI:
pydantic,pydantic-settings,PyYAML,typer,richRendering:
jinja2HTTP & retries:
httpx,tenacityData utilities:
python-dateutil,pandas,pyarrowLLM client:
openai
Credentials / secrets
OpenAI (for LLM features)
Provide an API key (recommended via environment variable):
OPENAI_API_KEY
Optional integration secrets (only if enabled)
Slack:
SLACK_BOT_TOKENSendGrid:
SENDGRID_API_KEYOpenSearch:
OPENSEARCH_API_KEY(or basic auth user/password, depending on config)Splunk HEC:
SPLUNK_HEC_TOKENMetabase/Postgres:
METABASE_PG_PASSWORD(or a password set in config, if allowed by your policy)
Optional integration dependencies
These are included as dependencies but are only used if the corresponding integration is enabled:
OpenSearch:
opensearch-pySlack:
slack_sdkSendGrid:
sendgridPostgres (Metabase warehouse):
psycopg[binary]
Recommended operational requirements
Treat inputs and outputs as sensitive (they may include hostnames, IPs, and vulnerability evidence).
Use environment variables for secrets and avoid committing credentials to Git.