lllllllamaCommunity·11 Agent Skills Included

ai-research-workflow-skills

Reproducible deep learning research reproduction, debugging, and exploration

Reproduces deep learning research repositories from README commands with auditable, evidence-linked reports. Diagnoses training failures, prepares environments and datasets, and runs conservative training or evaluation without silent code changes. Produces color-coded annotated READMEs and standardized evidence bundles so every result stays comparable, reversible, and verifiable.

515|11

↗

npx skills add lllllllama/ai-research-workflow-skills --all -g -y

Available:

Instructs the agent on how to route research requests to the correct reproduction, setup, training, debugging, or exploration skill while enforcing strict evidence and comparability rules.

All Skills in This Repository (11)

Pure Emerald Level Indicators

📦 In Repo

lllllllama

run-train

Execute deep learning training commands and record progress evidence in train_outputs.

minimal-run-and-audit

Capture and normalize execution evidence from short non-training commands into structured repro_outputs bundles.

ai-research-reproduction

Coordinate auditable reproduction workflows for AI research repositories.

paper-context-resolver

Resolve reproduction gaps between README guidance and primary papers.

safe-debug

Diagnose deep-learning errors from tracebacks and symptoms into structured diagnostic files.

analyze-project

Map entrypoints, model structure, and config relationships in deep learning repositories.

explore-run

Plan and summarize isolated exploratory deep learning experiment runs.

env-and-assets-bootstrap

Translate top-level environment files into a conda-first bootstrap flow with asset manifests.

ai-research-explore

Orchestrate auditable research campaigns with isolated branches and evaluation contracts.

explore-code

Isolate exploratory code changes on dedicated branches with traceable mappings.

repo-intake-and-plan

Scans READMEs and project files to extract documented commands and classify them into inference, evaluation, or training categories.

Community

Intermediate

Frequently Asked Questions

FAQPage Schema

How to install RigorPilot Skills?▼

Run `npx skills add lllllllama/ai-research-workflow-skills --all -g -y` in your terminal to install the full skill set globally.

How to reproduce a deep learning paper repository?▼

Use the ai-research-reproduction skill, which reads the README, picks the smallest documented inference or evaluation target, and writes a full evidence bundle to repro_outputs/.

What is the annotated README output?▼

It replays the target repository's README verbatim with color-coded, evidence-linked annotations showing exactly what ran, what was skipped, and what was blocked in each section.

Does it work with Claude Code and Codex?▼

Yes. All skills follow the open Agent Skills standard and run in Claude Code, Codex, Cursor, Gemini CLI, and other AGENTS.md-aware clients.

Can it explore new research ideas safely?▼

Yes, but only with your explicit authorization. The ai-research-explore skill runs candidate-only experiments on isolated branches with fair-comparison and rollback records.