chrisliu298
Community@chrisliu298 · Cupertino, California
Seedance @ ByteDance Seed; PhD Candidate @ UCSC; prev Post-Training @ Skywork AI
Agent Skills by chrisliu298
Showing 20 vetted skills indexed across 2 GitHub repositories.
chrisliu298publish-skill
Publish a local skill from dotfiles to a public GitHub repository.
chrisliu298autoresearch
Automate autonomous experimentation loops with versioned commits and metric evaluation.
chrisliu298atomic-push
Automate atomic commits and pushes from working tree changes.
chrisliu298prompt-engineer
Craft and refine prompts for Claude Code, GPT-5.4, and Codex models.
chrisliu298prism
Dispatch multiple agents to answer one question from diverse analytical lenses and synthesize perspectives.
chrisliu298xurl
Manage X posts, timelines, DMs, and API actions via the xurl CLI.
chrisliu298update-readme
Generate and validate README files from repository structure and manifests.
chrisliu298recall
Retrieve historical Claude Code sessions and Obsidian notes via QMD search.
chrisliu298citation-assistant
Automate citation lookup and verification for LaTeX documents across multiple sources.
chrisliu298push
Stage all changes, commit with repository conventions, and push to the remote.
chrisliu298rlm
Externalize AI agent state and orchestrate recursive subproblem processing across files.
chrisliu298deslop
Remove AI-generated commentary and dead code from git diffs.
chrisliu298arxiv-reader
Fetch arXiv papers by resolving IDs from URLs or bare IDs.
chrisliu298nanorepl
Decompose complex projects into progressive runnable layers with validation.
chrisliu298last-call
Identify session scope and review all changed files end-to-end.
chrisliu298Test-Driven Development
Enforce a test-first workflow requiring a failing test before production code.
chrisliu298interviewer
Facilitate structured mock interviews with adaptive questions and performance feedback.
chrisliu298chatgpt
Automates sending prompts to ChatGPT via Chrome and saves responses as Markdown files.
chrisliu298relay
Construct a Codex request, invoke Codex, and return the response via a single relay command.
chrisliu298karpathy-guidelines
Enforce behavioral guidelines for LLM-assisted coding to prevent overcomplication and dead code.