skills.rest
MichaelsEngineering
MichaelsEngineeringCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

gpd-map-research

Maps physics research projects into seven structured documents using parallel mapper agents.

Updated Mar 18, 2026
Search Tags:#documentation#parallel agents#codebase analysis#physics#gpd#formalism#research mapping
One-click install
npx skills add https://github.com/MichaelsEngineering/sentinel-multiscale-field-dynamics --skill gpd-map-research-michaelsengineering
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: gpd-map-research
Source: https://github.com/MichaelsEngineering/sentinel-multiscale-field-dynamics/tree/main/.agents/skills/gpd-map-research
Command: npx skills add https://github.com/MichaelsEngineering/sentinel-multiscale-field-dynamics --skill gpd-map-research-michaelsengineering

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Onboarding to or planning within an existing physics research project requires understanding scattered derivations, code, data, and conventions, which is slow and error-prone when done manually. ## Core Features & Use Cases - Parallel Research Mapping: Spawns four gpd-research-mapper agents covering theory, computation, methodology, and status, each writing documents directly to .gpd/research-map/. - Structured Output Documents: Produces FORMALISM.md, REFERENCES.md, ARCHITECTURE.md, STRUCTURE.md, CONVENTIONS.md, VALIDATION.md, and CONCERNS.md with file paths and line counts. - Safety and Refresh Handling: Detects existing research maps for refresh/update/skip, scans generated documents for leaked secrets, and commits results via the GPD CLI. - Use Case: Before initializing a new project phase, run the mapping to capture the current theoretical framework, solver pipeline, notation conventions, and open questions as reference material for planning. ## Quick Start Ask the AI to map the existing research project, optionally naming a focus area such as hamiltonian, numerics, or perturbation theory.

Frequently Asked Questions about gpd-map-research

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I map an existing physics research project?▼

Run the map-research workflow, which spawns four parallel mapper agents for theory, computation, methodology, and status. Each agent writes structured documents directly to .gpd/research-map/, and the orchestrator verifies line counts and commits the results.

What documents does the research mapping produce?▼

It produces seven Markdown files: FORMALISM.md, REFERENCES.md, ARCHITECTURE.md, STRUCTURE.md, CONVENTIONS.md, VALIDATION.md, and CONCERNS.md. Together they cover equations, literature anchors, solver pipelines, directory layout, notation, benchmarks, and open issues.

Can I update only part of an existing research map?▼

Yes. When .gpd/research-map/ already exists, the workflow offers Refresh, Update, or Skip. Choosing Update lets you name specific documents so only the relevant mapper agents rerun.

What happens if a mapper agent fails during mapping?▼

The workflow continues with the remaining agents and reports which focus areas failed. You can then retry the failed area or skip it, keeping a partial but usable research map.

Does the workflow check for leaked secrets before committing?▼

Yes. It greps the generated documents for common API key and token patterns. If matches are found, it pauses and waits for user confirmation before committing the research map.

When should I skip research mapping?▼

Skip it for brand new research directions with no prior work or trivial projects under five files, since there is nothing meaningful to map.

Related Skills

Explore More Education & Research Skills→
sidiangongyuansidiangongyuan

research-evidence

Automate academic literature search, citation verification, and claim evidence checking via arXiv API.

Community
Advanced
Chanw-researchChanw-research
38

experiment-plan

Convert research proposals into claim-driven experiment roadmaps with run orders and budgets.

Official
Advanced
dz306271098dz306271098

research-refine

Refine vague research directions into implementable plans through iterative GPT-5.4 reviews.

Community
Advanced
khapn-2933khapn-2933

tkm:scan-codebase

Scan codebases with parallel agents for file discovery and context gathering.

Community
Advanced