MohamedElashri

Create, review, render, validate, and locally test modern LHCb Run 3 Analysis Productions using info.yaml, lbexec entrypoints, Bookkeeping inputs, and lb-ap. Use for production job matrices, data or MC input queries, application versions, output registration, local tests, and preservation handoffs. Do not use for direct grid submission, Run 1/2 gaudirun.py productions, dataset discovery alone, or inventing working-group policy.

Create, review, run, and troubleshoot modern LHCb Run 3 DaVinci jobs using an options function, YAML data options, PyConf, and lbexec. Use for minimal DaVinci configurations, input-process and conditions setup, line filters, dry runs, and job artifact checks. Do not use for Run 1/2 gaudirun.py or configurable-style jobs, FunTuple variable design alone, data discovery, or grid production submission.

Design, review, and verify modern LHCb Run 3 FunTuple configurations and ROOT ntuple outputs. Use for decay fields, caret placement, ThOr functors, FunctorCollection composition, event variables, tuple naming, branch checks, and output inspection. Do not use for legacy DecayTreeTuple or LoKi TupleTools, general DaVinci environment setup, data discovery, or physics-variable choices without user requirements.

Create, review, and validate auditable LHCb Run 3 analysis specifications spanning physics objectives, datasets, software, selections, outputs, validation, provenance, open decisions, and preservation. Use for measurement plans, efficiency studies, control studies, analysis handoffs, and readiness reviews. Do not use to invent collaboration policy, approve physics choices, discover datasets by itself, or replace executable DaVinci and production validation.

Find, compare, record, and validate LHCb Run 3 data or simulation datasets using Bookkeeping, LHCbDIRAC, processing metadata, LFNs, and access checks. Use for event types, campaigns, magnet polarity, streams, HLT2 or Sprucing outputs, file manifests, and dataset provenance. Do not use for generic web search, physics-literature discovery, direct production submission, or choosing undocumented sample metadata.

Select, construct, and verify a current LHCb Run 3 released or development software environment. Use for lb-run, lbexec, lb-dev, platform selection, DaVinci version selection, CVMFS availability, or diagnosing an LHCb application environment. Do not use for generic Python environments or legacy Run 1/2 setup unless the user explicitly requests legacy support.

Use when preparing postdoc applications, research statements, cover letters, CV bullets, interview answers, short talks, emails, or role-fit analysis for HEP, LHCb, trigger/software, data science, or research-software roles.

Use when reviewing code, tests, scripts, notebooks, configuration, diffs, or physics-analysis workflows for correctness, maintainability, performance, reproducibility, or scientific validity.

Use when reviewing, optimizing, or debugging CUDA performance, especially LHCb Allen trigger software or GPU-resident reconstruction code.

Use when implementing features, fixing bugs, refactoring, updating tests/docs/config, or making any non-trivial repository change.

Search, browse, and analyze CERN GitLab repositories for HEP code, documentation, and examples. Use this skill whenever the user mentions CERN GitLab, LHCb/ATLAS/CMS code, HEP software stacks, or wants to find code patterns, inspect build systems, read READMEs, search issues, or explore releases in gitlab.cern.ch projects. Also trigger when the user references specific CERN projects like DaVinci, Allen, Athena, Gauss, Moore, Boole, or any gitlab.cern.ch URL.

Analyze CERN ROOT files using the root-cli command-line tool. Use this skill for any task involving ROOT files (.root), TTree/RNTuple inspection, branch reading, histogram computation, kinematic reconstruction, fitting, correlation analysis, or exporting HEP data. Trigger whenever the user mentions ROOT files, TTrees, RNTuples, for any task involving ROOT files (.root), TTree inspection, branch reading, histogram computation, kinematic reconstruction, fitting, correlation analysis, or exporting HEP data. Trigger whenever the user mentions ROOT files, TTrees, branches, histograms, invariant mass, MET, particle physics data, or asks to inspect/plot/fit/export data from .root files. Prefer root-cli over PyROOT/ uproot scripting for interactive analysis — it saves thousands of tokens and provides a clean, composable command surface. Also trigger when the user asks to use root-mcp in CLI mode.