dbugs CLI

A command-line client for the dbugs vulnerability database by Positive Technologies. Query vulnerabilities, trends, references, news, and researchers from your terminal.

Output is rendered as Rich tables by default, or as raw API JSON with --json (pipe-friendly for jq). The tool is read-only and needs no API key or login.

Commands

Command	Purpose
stats	Global database totals
vulns	Search / list vulnerabilities (filter, sort, export)
vuln <id>	Full detail for one vulnerability, incl. references
trends	Trending vulnerabilities (filtered/sorted client-side)
trend <id>	Social-media posts driving a trend
news	List / filter security news (filter, export)
news-item <slug>	One news article
suggest products|vendors [PATTERN]	Discover valid --product/--vendor values
researcher <name>	Researcher profile and stats

Install

pipx install .
# or, for development:
pip install -e ".[dev]"

Usage

dbugs stats                                   # database totals
dbugs vulns --fts apache --limit 10           # full-text search
dbugs vulns --vendor microsoft --severity CRITICAL --has-exploit --sort score
dbugs vulns --min-score 9 --since 2026-07-01  # score + date filters
dbugs vulns --vendor microsoft --export vulns.jsonl   # all matches -> JSONL
dbugs news --product Wordpress --export news.json      # all matches -> JSON
dbugs vuln PT-2026-61063                       # full detail incl. references
dbugs vuln PT-2026-61063 --source Exploit --source Note   # filter references by source
dbugs news                                      # latest security news
dbugs news --product Wordpress --vendor Microsoft --fts rce
dbugs news --cve CVE-2026-63030 --since 2026-07-01
dbugs trends                                    # all 30 trending vulnerabilities
dbugs trends --min-score 9 --severity CRITICAL --sort posts --limit 10
dbugs suggest products word                     # discover valid --product values
dbugs suggest vendors                           # popular vendors (no pattern)
dbugs trend PT-2026-53941                       # social posts behind a trend
dbugs news-item <slug>                          # one article
dbugs researcher <name>                         # researcher profile

# Machine-readable output for any command:
dbugs --json vulns --fts apache | jq '.rows[].vulner_id'

Filtering news

dbugs news supports server-side filters: --fts, --product, --vendor, --researcher, --cve, --category (all repeatable except --fts), and --since/--until (published date range). Use dbugs suggest products|vendors [PATTERN] to discover valid product/vendor values.

Filtering trends

The service returns a fixed set of 30 trending vulnerabilities. dbugs trends filters and sorts them locally: --min-score, --severity, --min-posts, --sort score|posts, --asc, --limit. Under --json the filtered set is emitted.

Filtering vulnerability references

dbugs vuln <id> accepts --source (repeatable) to keep only references whose source matches — exact, case-insensitive. Common sources are Exploit, Note, and Vendor Advisory (quote multi-word values). The filter applies to both the table and --json output.

dbugs vuln PT-2026-61063 --source Exploit
dbugs --json vuln PT-2026-61063 --source "Vendor Advisory" | jq '.references'

Exporting full results

dbugs vulns and dbugs news accept --export PATH to write the entire filtered result set to a file, auto-paginating through every page (the API's own paging cap no longer applies). While exporting, --limit/--page are ignored and normal table/--json output is suppressed; progress is printed to stderr.

The format is inferred from the file extension — .jsonl writes one JSON object per line (streamed, best for large results), any other extension writes a single {"count": N, "rows": [...]} document. Override with --format json|jsonl.

dbugs vulns --vendor microsoft --severity CRITICAL --export out.jsonl
dbugs news --product Wordpress --export news.json --format jsonl

Global options

• --json — emit raw API JSON instead of tables.
• --locale en — response locale.
• --timeout 30 — HTTP timeout in seconds.

Notes

The service sits behind anti-bot protection; the client sends the required browser headers automatically. No API key or login is needed. The tool is read-only.

Development

pip install -e ".[dev]"   # install with dev deps (pytest, respx)
pytest                    # unit tests (network mocked)
DBUGS_LIVE=1 pytest       # also run tests against the live API

Project layout

dbugs_cli/
  cli.py          Typer commands, argument parsing, error boundary
  client.py       the only module that touches HTTP (+ browser headers)
  models.py       dataclasses parsed from API responses (each keeps .raw)
  transform.py    client-side filter/sort (used for trends)
  export.py       auto-paginating full-result writer (JSON/JSONL)
  formatters.py   pure model -> Rich renderable functions
tests/            one file per module; HTTP mocked with respx
docs/superpowers/ design specs and implementation plans

Modules are strictly layered: cli → client → models, with transform, export, and formatters as leaf helpers. See CLAUDE.md for the conventions to follow when extending the tool.