{{ ai_component_prompts|length }} component prompts. Start here — if the LLM says not_affected, all findings on that component are likely FP.
{{ p.prompt }}
{{ ai_triage_prompts|length }} finding prompts. Only needed for findings on components not already ruled out above.
{% for p in ai_triage_prompts[:20] %}{{ p.prompt }}
Showing first 20 of {{ ai_triage_prompts|length }}. Full set in the _prompts.md file.
{% endif %}| CVE | Component | Severity | Confidence | Signals | Reason / AI Rationale | Action |
|---|---|---|---|---|---|---|
| {{ row.cve_id or row.finding_id or '—' }} | {{ row.component_name or '—' }} {{ row.component_version or '' }} | {{ row.severity or 'UNKNOWN' }} | {{ row.fp_confidence or '—' }} | {% if row.fp_signals %} {% for sig in row.fp_signals.split(',') if row.fp_signals is string %} {{ sig.strip() }} {% endfor %} {% if row.fp_signals is not string %} {% for sig in row.fp_signals %} {{ sig }} {% endfor %} {% endif %} {% endif %} |
{{ row.primary_reason or '—' }}
{% if row.ai_rationale %}
AI details
Verdict: {{ row.ai_verdict }}
{{ row.ai_rationale }} |
{{ row.recommended_action or 'Review' }} |
No false positive candidates detected. {% if not summary or summary.ai_detections == 0 %}Try running with --ai for AI-based applicability analysis.{% endif %}
| Signal | Count |
|---|---|
| {{ sig }} | {{ count }} |