llnl/dftracer-agents

dftracer-annotate-general

Language-agnostic dftracer annotation rules — Rule 0 (what to skip), coverage, error paths, backend priority order, and wrapper handling

dftracer-annotate-python

Python annotation rules for dftracer — decorator usage, initialize/finalize, comp types, class methods, and quick checklist

dftracer-annotation-lessons

Lessons learned from dftracer annotation sessions — real errors, root causes, and exact fixes. Loaded by each per-file annotation subagent at startup. Updated by the pipeline recipe after every session (Step 8).

dftracer-communication-optimization

Communication-component (MPI/collective/network) bottleneck-to-optimization mappings, papers, and L1/L2/L3 strategies for the dftracer optimization pipeline

dftracer-compute-optimization

Compute-component bottleneck-to-optimization mappings, papers, and L1/L2/L3 strategies for the dftracer optimization pipeline

dftracer-diagnoser

Diagnose traces and convert symptoms into actionable bottlenecks with MCP tools.

dftracer-memory-optimization

Memory-component bottleneck-to-optimization mappings, papers, and L1/L2/L3 strategies for the dftracer optimization pipeline

dftracer-optimization-kb

Cross-session, citation-backed knowledge base of every MEASURED dftracer optimization, partitioned into system-centric (L3), software-centric (L2), and workload-centric (L1) findings. Load this FIRST in any optimization loop to avoid re-deriving what is…

dftracer-overlap-analysis

How to correctly measure temporal overlap between two trace categories (e.g. compute vs data-loading, data vs communication) across MULTIPLE PROCESSES — a dynamic-interval, event-splitting algorithm that avoids both the "same-pid sweep-line gives a false 0%"…

dftracer-privacy-guard

MANDATORY end-of-session privacy validation. Scans everything the session added to the git-tracked trees (memory, lessons, skills, agent definitions) for identifying content — usernames, absolute user paths, emails, flux job ids, session UUIDs, node hostnames…

software-autotools

Autotools (autoconf/automake/libtool) build system detection, bootstrap/configure/build/install flow via session_configure/session_build_install/session_build_annotated, PKG_CONFIG_PATH/CPPFLAGS/LDFLAGS dftracer injection, and per-system/per-dependency lesson…

software-cmake

CMake build system detection, configure/build/install flow via session_configure/session_build_install/session_build_annotated, CMAKE_PREFIX_PATH dftracer injection, and per-system/per-dependency lesson tagging. Load this skill for any project whose top level…

software-flux-fiction

Flux-Fiction-specific knowledge: build system (Meson), dependency quirks, C99 standard handling on Cray clang, and the priority annotation focus (job-emulation lifecycle using the simulated/faketime clock). Flux-Fiction is software being…

software-meson

Meson build system detection, configure/build/install flow, dependency() resolution (pkg-config vs workspace/prefix fallbacks), and meson-python packaging. Load this skill for any project whose top level is meson.build / meson.options.

software-pecan

PECAN (Pose Classification, part of PECAN/MILAN pose-classification/binding-affinity codebase) — PyTorch+PyTorch-Geometric EGNN training on HDF5 shard data. Build, annotation status, dataset access, and measured dftracer optimization findings. Load this skill…

software-pip

pip / PEP 517-518 build-backend detection (setuptools, mesonpy, hatchling, ...), editable vs non-editable install choice, and venv/module-load ordering on HPC systems. Load this skill for any project installed via `pip install` off a pyproject.toml.

system-tuolumne-spindle

Spindle scalable library/Python-module loader on Tuolumne (LC toss4/Cray) — what it is, how it's already wired into flux by default, how to tune its level for Python/DL-heavy workloads, and how to verify it's active. Load this skill for any job launch-time /…

tools-dftracer

The dftracer PyPI package itself (C core + CMake build) — package layout, dependency graph (dftracer-utils, pybind11, pydftracer), install sources (PyPI release vs GitHub develop via pip), and session-local vs shared-venv install rules. Load this skill before…

tools-dftracer-utils

dftracer-utils — the Python binding package for the trace-utils CLI (split/merge/compact/reader/event_count on .pfw/.pfw.gz traces). Install verification and its relationship to the dftracer/pydftracer package trio. Load this skill before any trace…

tools-pydftracer

pydftracer — the Python annotation API package (dftracer.python module, dft_fn/dftracer_fn decorators, initialize_log/finalize). Install verification, common ModuleNotFoundError, and the API surface every python_annotate_file-generated file depends on. Load…

dftracer-io-optimization

Key literature, bottleneck-to-optimization mappings, and strategies for the dftracer I/O optimization pipeline

workload-vpic-kokkos

VPIC-Kokkos-specific knowledge: build/annotation quirks, the two-stage deck-compile pattern, deck sizing for smoke vs validation runs, and the measured MPI-communication-bound diagnosis on Tuolumne. Load this skill whenever working with vpic-kokkos…

dftracer-build-dftracer

Install and build dftracer into the session using the MCP install tool and record tool pitfalls early.

dftracer-annotate-c

Complete C annotation rules for dftracer — START/END placement, comp types, linker setup, coverage verification, and the full checklist

software-mpifileutils

mpifileutils (dcp/dsync/drm/dwalk/dcmp/dtar/dchmod/dstripe/etc.) — parallel, MPI-based replacements for serial Linux filesystem utilities (cp/rsync/rm/ find/diff/tar/chmod) on HPC parallel filesystems. Load this skill whenever a filesystem operation (copy,…

workload-h5bench

H5Bench-specific knowledge: CMake build quirks, annotation edge cases (assert/else-if brace insertion), INI config format, dftracer integration, and Cray HDF5 compatibility. Load this skill whenever working with h5bench.

workload-ior

IOR-specific knowledge: build quirks, annotation pitfalls, ROMIO tuning on VAST NVMe storage, and optimal dftracer run configuration for IOR 4.0.0. Load this skill whenever working with IOR.

dftracer-cheatsheet

Critical annotation rules, corner cases, and top-10 known mistakes for dftracer — read this before every annotation session

dftracer-install

Install and privilege rules for dftracer sessions — never use sudo, always install to userspace paths; autotools pkg-config integration

dftracer-pipeline

Interactive dftracer annotation pipeline. Clones the app, builds it, annotates all source files using the clang-based MCP annotation tools (clang_annotate_project / clang_annotate_file + clang_syntax_check + clang_lint_annotations), builds the annotated…

dftracer-preload-run

Reference for running applications with dftracer in PRELOAD or HYBRID mode. Covers correct environment variable setup, DATA_DIR rules, LD_PRELOAD wiring, MPI forwarding, and trace verification. Use before any session_run_with_dftracer call or when debugging…

dftracer-trace-utils

TOP PRIORITY — always use dftracer MCP utils tools for any trace work; never use raw bash/python/gzip scripts to read or process .pfw/.pfw.gz files

software-mpi

MPI and MPI-IO (ROMIO) optimization strategies, Cray MPICH environment variable wiring, flux proxy env propagation rules, and dftracer MPI tracing requirements. Load this skill for any MPI or collective I/O work.

workload-scaffold

ScaFFold (LBANN PyTorch benchmark, 3D U-Net on procedurally-generated fractals). Python/PyTorch DDP app running on Tuolumne MI300A. Environment setup, shared venv pattern for AI/ML workloads, MPI4PY Cray linkage, ROCm paths.

dftracer-lessons

How to use and update the lessons-learned cache at .goose/lessons-learned.md — proactive lookup before any build/install/annotate step

dftracer-profiling

MANDATORY pipeline self-profiling via MLflow. Every dftracer agent brackets its work with profile_step_begin / profile_step_end so the session records what each step cost in wall time, tokens, dollars, and retries. The main thread binds the session once with…

workload-flashx

Flash-X-specific knowledge for the dftracer pipeline — build system (Python setup + GNU Make), Sedov 3D test problem, dangling symlinks in shallow clones, HDF5/flash.par pitfalls. Load this skill whenever working with Flash-X.

dftracer-annotate-process

The complete dftracer manual annotation process — 5 passes, coverage requirements, annotation_logs format, build prep, and mandatory Pass 5 review

dftracer-context-economy

Context/token economy for every dftracer agent. Use the graph tools (graph_query / graph_ensure) to LOCATE code and compute change blast-radius instead of reading whole files into context. Load this before any tool call, annotation, optimization, or pipeline…

dftracer-ml-annotate

Annotate ML/DL Python workloads end-to-end with dftracer AI/ML region decorators. Detects frameworks and ROCm/HIP requirements, installs dftracer with correct flags, annotates all Python files, builds/runs smoke test, collects traces, and auto-updates the…