DigitalOcean Referral Badge
cloud1
cloud2
cloud3
cloud4
cloud5
cloud6
← Back to catalog
code hdrinj

Header injection attempts

Syndu annotates all incoming traffic and extracts behavioral signals that help explain intent. This page defines the hdrinj signal — what it means, how to interpret it, and how it will later connect to live evidence across IPs, subnets, organizations, ISPs, countries, and cities.

Signal gist Input patterns suggest attempts to manipulate headers or downstream header parsing.
Read the definition → View live signal sections →

Definition

Canonical reference for hdrinj behavior.
Catalog code
hdrinj
Display name
Header injection attempts
How to read this signal
This annotator represents a behavioral pattern, not a claim of identity. It’s designed to help you understand why certain traffic looks suspicious, automated, probing, or exploit-oriented — and to support consistent reporting across the Syndu system.
Explanation
Flags indicators consistent with header injection or header-structure manipulation, such as suspicious delimiter/control-character patterns and payloads aimed at confusing proxies or application parsing. This can be relevant to cache/proxy edge behavior and request smuggling-adjacent classes. Present evidence carefully (avoid excessive payload disclosure) and interpret with endpoint and repetition context.

Live sections

These panels will be wired to real metrics, enrichment context, and drill-down links.
Signal footprint over time
Rolling volume, bursts, first/last seen, and time-window slices (e.g. last hour/day/week). This will help separate chronic background noise from active campaigns.
Coming next: time series + burst markers
Top affected entities
Links to the entities where hdrinj is most present: IPs, subnets, organizations/ASNs, ISPs, and geographies — with “why” context.
Coming next: entity leaderboards + drill-down
Enrichment context
How enrichment affects interpretation: known crawlers, monitored ranges, trusted scanners, or policy exceptions. This is where “benign but noisy” gets separated from “unknown and risky.”
Coming next: enrichment flags + allowlist context