Abstract
With the growing number of gamma-ray monitors in operation, research teams have adopted joint operation strategies to improve scientific efficiency. The GECAM-HXMT-SVOM (GHS) constellation collaboration serves as a successful precedent, yet joint duty presents challenges to Burst Advocates (BAs), including a high volume of triggers and system incompatibilities. To address these requirements in multi-messenger astronomy, we developed the BA’s Rapid Evaluation and Analysis Kit for Formulating Alerts and Summary Tools (BREAKFAST). This framework incorporates automated and semi-automated pipelines to process prompt emission triggers while maintaining compatibility for future missions. Beyond processing, BREAKFAST filters high-value targets and guides follow-up telescopes by predicting X-ray afterglow brightness based on prompt emission properties. The framework's effectiveness has been demonstrated in recent campaigns, highlighting its role in the discovery of peculiar transients.
Introduction
We are currently in the golden age of Time-Domain and Multi-Messenger Astronomy (TDAMM). However, this success has created a "data deluge." Multiple $\gamma$-ray monitors are constantly triggering on events like Gamma-Ray Bursts (GRBs) and Magnetar X-ray Bursts (MXBs). Traditionally, each mission—such as Fermi, Swift, or GECAM—has maintained its own team of Burst Advocates (BAs) and proprietary data formats.
As research teams move toward "joint duty"—where a single BA monitors multiple satellites—the friction of switching between incompatible systems has become a significant bottleneck. BREAKFAST was designed to unify these disparate data streams into a single, cohesive framework that automates standard analysis and provides actionable intelligence for follow-up observations.
The BREAKFAST Framework
BREAKFAST represents a shift from mission-centric to constellation-centric operations, treating a fleet of satellites as a single instrument. The framework is structured into three primary modules:
1. Trigger Module
This module polls internal servers (GECAM, SVOM) and external networks (GCN, AstroBurstHub). A novel feature of this module is the integration of Large Language Models (LLMs) to parse human-written alerts, such as Astronomer's Telegrams (ATels), into structured database entries. This solves a long-standing interoperability headache in the astronomical community.
2. Analysis Module
The "engine room" of the framework invokes specialized pipelines (e.g., ETJASMIN and ELISA) to calculate critical burst parameters. These include $T_{90}$ (burst duration), spectral lags, and hardness ratios. The module also performs automated spectral fitting using Powerlaw and Band models to characterize the energy distribution of the prompt emission.
3. Interactive Module
A Flask-based web interface and terminal scripts allow BAs to manually refine results, update localizations, and trigger follow-up requests. This human-in-the-loop design ensures that automated results are verified before being disseminated to the wider community.
Predictive Analytics for Follow-up Guidance
A core contribution of BREAKFAST is its ability to bridge the gap between prompt emission and afterglow observations. Using the "closure relation" between prompt and afterglow phases, the framework implements a power-law relation ($L_X/E_{iso} \propto t_{rest}^{-1.23}$) to estimate X-ray afterglow brightness. By predicting whether a burst will remain detectable by X-ray telescopes like the Einstein Probe (EP) hours after the initial trigger, BREAKFAST helps prioritize limited observation time for high-value targets, such as potential kilonovae or high-redshift candidates.
Operational Results and Case Studies
BREAKFAST has already demonstrated its utility in real-world observational campaigns:
- GCN Production: The framework has been used to generate over 100 GCN Circulars, significantly reducing the manual workload for BAs.
- GRB 250516B: BREAKFAST successfully processed this bright burst, providing rapid IPN-compatible localization and spectral analysis across both GECAM and HXMT data.
- Predictive Success (GRB 250919A): The framework predicted an X-ray afterglow flux $> 10^{-10}$ erg·cm⁻²·s⁻¹ at two hours post-trigger. This prediction was confirmed by the Einstein Probe (EP/WXT), validating the framework's guidance capabilities.
- High-Redshift Identification: The system assisted in identifying GRB 250314A as a high-redshift candidate ($z \sim 7.3$), a discovery vital for studying the early universe.
Future Directions
While BREAKFAST has streamlined joint duty, future work will focus on automating joint multi-instrument spectral fitting and implementing machine learning techniques to better distinguish GRBs from solar flares or other noise. Plans are also underway to integrate data from the POLAR-2 mission and the GRID satellite constellation, eventually opening the platform to the broader TDAMM community via the AstroBurstHub.
Summary updated