Skip to content
HUNTERPIGGY
Signal → Trajectory
Explainable · local demo · human-reviewed

HUNTERPIGGY

Detect the signal. Reconstruct the trajectory.

HunterPiggy finds connected warning signals in messages and systems, explains why they matter, and reconstructs the mechanisms that may have produced them.

Investigation leads, not verdicts. Human review is required.

ExplainableDeterministicHuman-reviewed
HunterPiggy mascot inspecting a suspicious link with a magnifying glass, connected to a neon signal network.
case://sus.link/7xR9…◉ live demo
Signal to trajectory pipeline: input, detect, map, simulate, human review. Markers are not verdicts.
  1. 01
    Input
    Prepared demo or your own text.
  2. 02
    Detect
    Correlated markers, not isolated flags.
  3. 03
    Map
    Origin, context, behavior, intent, pattern, linkage.
  4. 04
    Simulate
    Possible trajectories — labelled as hypotheses.
  5. 05
    Human review
    Every finding shows its evidence.
Choose a demo

No setup required. Demo inputs are editable.

Try it in 20 seconds:1Load a demo2Edit or keep it3Run analysis4Open evidence
Signal input
case: scam-parcelnova-001
Scam & Manipulation Signal Review
548 characters · edits change the next analysis
Analysis

No report yet

Ready to investigate. Load a demo or paste your own text, then run the analysis to reveal connected markers, evidence and possible trajectories.
Review boundaries
HunterPiggy analyses supplied content and deterministic patterns.
A marker is an entry point for investigation, not evidence of wrongdoing.
A possible trajectory is a hypothesis, not a prediction or causal proof.
Context, satire, translation and incomplete evidence can produce false positives.
Results require human review.
Do not use the report as the sole basis for punitive, legal, financial or safety-critical action.
How this was built
Before / Now diagram: HunterPiggy graduated from an internal signal router inside Python Zero to a standalone package with causal trajectories, PIC–Boson adapter and audit trail, built during OpenAI Build Week.

Before Build Week: HunterPiggy existed inside the broader Python Zero / Kai256 research system as a transparent signal router that detected and prioritised anomaly markers.

Built during Build Week: standalone HunterPiggy product and repository; signal-to-trajectory workflow; explainable evidence presentation; PIC–Boson institutional layer; Growth and Shadow trajectory simulation; counter-hypotheses; human-review boundaries; prepared demonstrations and documentation.

GPT-5.6

Used for product architecture, conceptual synthesis, safety boundaries, PIC–Boson mapping and demonstration design.

Codex

Used to inspect and structure the codebase, implement and refine modules, improve tests and documentation, verify the standalone build and clearly separate prior work from the Build Week contribution.