prml-vslam

Use when writing, revising, or reviewing PRML VSLAM final report or slide deliverables under docs/report or docs/slides, including Typst figures, tables, citations, bibliography references, scientific claims, VSLAM metrics, alignment, method or dataset framing, protocol text, limitations, result statements, and final presentation prose.

Create, update, or publish GitHub pull requests with a reviewer-first structure. Use when Codex needs to draft or revise a PR title/body, open a PR from a prepared branch, or reshape an existing PR into a summary paragraph, focused verification list, compact work-package map, and detailed work-package sections instead of a raw changelog.

Create, de-duplicate, structure, sync, and resolve GitHub issues. Use when Codex needs to open a new GitHub issue, rewrite a vague issue into a template-backed body with acceptance criteria, mirror issues between GitHub and another backlog, or close the loop from issue to linked pull request.

Write and refactor concise, contract-focused Python docstrings for modules, public classes, functions, methods, protocols, config models, DTOs, wrappers, and streaming or session APIs. Use when Python docstrings are missing, sparse, misleading, or need better cross-references, examples, units, shapes, lifecycle notes, or boundary semantics.

Use when working with the repo-local MemPalace setup for PRML VSLAM, especially to refresh or search mined docs and Codex chat histories, inspect palace status, or generate wake-up context from the repo-local palace.

Use when the user asks to reduce redundancy, boilerplate, dead code, unused symbols, unused config fields, overcomplication, or Python source LOC through behavior-preserving simplification, pruning, or refactoring in this repo.

Use when working in a repository with graphify-out/, especially before answering architecture or codebase questions and after code edits that should refresh graphify artifacts.

Use when creating, reviewing, or fixing Python Rerun integrations for SLAM, RGB-D, reconstruction, and spatial-vision workflows, especially when work involves transforms and frame conventions, pinhole cameras, depth images, point clouds, timelines, blueprints, or comparing a local integration against official Rerun examples or the repo-local ViSTA-SLAM reference.

Use when working with PRML VSLAM's internal agent memory (`.agents/AGENTS_INTERNAL_DB.md`, `.agents/issues.toml`, `.agents/todos.toml`, `.agents/refactors.toml`, `.agents/resolved.toml`) or triaging, resolving, and maintaining the backlog with `make agents-db`.

Explains Streamlit's internal architecture including backend runtime, frontend rendering, and WebSocket communication. Use when debugging cross-layer issues, understanding how features work end-to-end, planning architectural changes, or onboarding to the codebase. Covers ForwardMsg/BackMsg protocol, script rerun model, element tree, widget state management, and more.