peabody124
- 22 skills
- 0 followers
- 1 week ago last updated
- ▌ Scaffold Repo · peabody124Initialize a new Python repository with correct structure following RAE guidelines. Use when the user says "new repo", "new project", "initialize", "scaffold", "create a project", "set up a repo", or "start a new package". Creates src/ layout, pyproject.toml, tests, .gitignore, and optionally a devcontainer.
- ▌ Pose Datajoint · peabody124Use when writing Python code to query biomechanics DataJoint tables - counting videos/sessions, filtering by video_project or participant_id/subject_id, fetching keypoints or kinematic reconstructions, synchronizing keypoints with qpos, understanding Session-Video relationships for both multi-camera and monocular pipelines
- ▌ Gait Lab Dataset · peabody124Query and analyze clinical gait lab (marker-based motion capture) data. Use when asked about c3d files, gait lab sessions, mocap markers, MocapProjection synchronization, video-mocap temporal alignment, GaitLab-MMC trial matching, gait events, gait parameters, forceplate/EMG analog data, or comparing marker-based and markerless motion capture systems.
- ▌
- ▌ Consult Guidelines · peabody124Review relevant coding guidelines before starting a task. Use when the user says "consult guidelines", "check standards", "what are the rules for", "review guidelines", or when starting Python, git, refactoring, debugging, or code review work. Also useful when asking "what line length", "how should I format", "what's the testing policy".
- ▌ Enforce Guidelines · peabody124 bundleMandatory skill that ensures all work follows RAE guidelines. Activates automatically before any code task.
- ▌ Jax Jit And Fusion · peabody124Find and fix JIT/fusion-hygiene misses in JAX/Equinox code: expensive work running eager (un-jitted), under-fused (many tiny jits, or per-element Python loops calling un-jitted forwards), or missing `eqx.filter_jit` on a hot path. Use when "every step is slow", "GPU sits idle", "low GPU utilization", a model "runs eager" / "not jitted" / "per-crop" / "per-frame in a Python loop", or when you want to "fuse" ops, choose a "jit boundary", set up `filter_jit`, `static_argnums` / `static_argnames`, `donate_argnums` / buffer donation, or decide `lax.scan` vs an unrolled loop. Also use proactively before shipping a hot inference path, or when benchmarking eager-vs-jit (warm + block_until_ready). Complements `jax-memory-and-retracing` — fusion buys speed but can raise peak memory and trigger retraces; check both.
- ▌ Efficient Rendering · peabody124Use when building or speeding up visualization that renders 3D meshes, point clouds, skeletons/keypoints, or overlays onto video — especially offscreen/headless (EGL) rendering, differentiable rasterization in JAX, compositing renders over real camera footage, or anything that must align with a calibrated camera (intrinsics/extrinsics/distortion). Also use for the symptoms — "rendering is slow", "llvmpipe" / CPU GL fallback, "OffscreenRenderer slow", render is "upside-down" or "mirrored", overlay "doesn't line up with the video", blank/clipped renders, or an OOM in a rasterizer. Covers the OpenCV↔GL camera conventions, the two distortion strategies (undistort-the-image vs bake-distortion-into-vertices), GPU device probing and context reuse, chunked/checkpointed rasterizers, and the video-overlay pipeline. Pairs with the camera-model skill for the geometry API.
- ▌ Golden Test Harness · peabody124Set up self-generating golden/parity integration tests where a config-driven test harness generates the golden data on first run and the golden files are .gitignored, not committed. Use when adding parity tests, comparison tests, integration tests, or golden fixtures; when deciding whether to commit golden data; or when the user says "generate goldens", "golden test", "parity test", "regenerate golden", "integration test data", or "should I commit goldens". Covers config-points-to-DataJoint-trial, checked-in fixture images, pytest flags, and the rare cases where vendoring goldens is acceptable.
- ▌ Github Activity Review · peabody124Use when the user asks for a periodic review of their GitHub activity, wants a summary of what they've been working on, asks "what have I done recently", "review my commits", "summarize my GitHub activity", or wants a cross-repository progress report. Also use when asked to review team activity, lab progress, or multi-repo development summaries. Defaults to 2-week window. Creates thematic report in scratch/.
- ▌ Bead Driven Development · peabody124 bundleThis skill should be used when the user asks to "execute plan with beads", "implement with persistent tracking", "bead-driven execution", "run plan with subagents", "track tasks across sessions", "use beads for planning", or wants planning + subagents + beads + code review unified. Orchestrates writing-plans, executing-plans, and investigation skills with beads integration for persistent cross-session tracking.
- ▌ Jax Memory And Retracing · peabody124Diagnose and prevent GPU/host memory OOMs, memory leaks, and JIT retracing in JAX/Equinox projects. Use when a process is "killed" with no traceback, when you see "OOM", "out of memory", "GPU OOM", "RESOURCE_EXHAUSTED", "memory leak", "host RAM grows", "retracing", "recompiling every step", "every step is slow", "silent crash / no traceback", "cudnn mismatch", or "cache deserialize error". Also use proactively when writing per-frame/per-step loops around jit'd functions, accumulating arrays over a long sequence, enabling a persistent compilation cache, or choosing float precision for a memory-constrained run.
- ▌ Fetching Synchronized Data · peabody124Use when fetching 2D keypoints synchronized with kinematic reconstruction qpos, aligning video frames with reconstruction frames, working with fetch_keypoints, handling camera parameter ordering between Calibration and SingleCameraVideo, or debugging temporal misalignment between observations and poses
- ▌ Semi Formal Code Reasoning · peabody124Structured semi-formal reasoning templates for agentic code analysis. Use this skill whenever Claude is asked to reason deeply about code semantics without executing it — including: comparing two patches or code changes for equivalence, reviewing pull requests or diffs for correctness, localizing bugs or faults from failing tests, answering questions about how code behaves across a repository, verifying refactors preserve behavior, or any task where Claude must trace execution paths through multi-file codebases. Trigger on phrases like "are these patches equivalent", "review this diff", "find the bug", "why does this test fail", "does this refactor change behavior", "trace through this code", "what does this code actually do", or any code analysis task spanning multiple files where getting the reasoning right matters. Also trigger when the user asks Claude to act as a code verifier, reward model, or static analyzer. Do NOT use for simple syntax questions, formatting, or code generation tasks where execution t
- ▌ Datajoint Biomechanics Schema · peabody124Use when working with biomechanics DataJoint pipeline - querying KinematicReconstruction, understanding Video-Session relationships, using bridging algorithms, debugging schema issues across PosePipeline/MultiCameraTracking/BodyModels, or working with method 137 results
- ▌ Deslop · peabody124Clean up AI-generated slop from code changes. Use when the user says "deslop", "clean up code", "remove slop", "review for AI artifacts", "check for anti-patterns in diff", or before committing to ensure code quality. Also activates when reviewing staged changes or preparing a commit.
- ▌ Audit Repo · peabody124Audit the current repository against RAE standards. Use when the user says "audit repo", "check repo standards", "sync with RAE", "check for drift", "compare against template", or when you want to verify a repo follows RAE conventions. Compares pyproject.toml, CLAUDE.md, ruff config, pytest/coverage config, and dev dependencies against the canonical RAE template.
- ▌ Excalidraw · peabody124Use when asked to create diagrams, flowcharts, architecture diagrams, wireframes, visual drawings, or .excalidraw files, or to assemble a composite scientific overview figure (rendered chart/3D panels embedded under a vector annotation layer). Generates valid Excalidraw JSON, embeds raster panels, renders to PNG/SVG, and visually verifies the result.
- ▌ Jax Config · peabody124 bundleSet up JAX/Equinox-compatible configuration for a project. Use when the user says "set up config", "add configuration", "make a config", "jax config", or when creating dataclass configs that will be passed through JAX-traced code. Also use when debugging recompilation issues or pytree registration problems.
- ▌ Camera Model · peabody124Use when working with camera projections, intrinsics, extrinsics, triangulation, reprojection, coordinate transforms between world/camera frames, or any code that touches camera_params dicts. Covers unit conventions (mm vs meters), the OpenCV world-to-camera extrinsic convention, and the required API functions from camera.py and camera_adapter.py.
- ▌ Gait Metrics · peabody124Use when analyzing gait data, computing spatiotemporal metrics, finding valid walking segments, calculating step length/cadence/velocity, joint ROM per gait cycle, gait phase percentages, asymmetry indices, Gait Deviation Index (GDI), GaitAnalytics table queries, or normative gait comparisons from kinematic reconstructions
- ▌ Investigation · peabody124Scaffolds a structured investigation in scratch/ for empirical research and documentation. Use when the user says "start an investigation" or wants to: trace code paths or data flow ("trace from X to Y", "what touches X", "follow the wiring"), document system architecture comprehensively ("document how the system works", "archeology"), investigate bugs ("figure out why X happens"), explore technical feasibility ("can we do X?"), or explore design options ("explore the API", "gather context", "design alternatives"). Creates dated folder with README. NOT for simple code questions or single-file searches.