chrisliu298

@chrisliu298 source repo

34 skills
0 followers
5 hours ago last updated

▌
Chatgpt 2 · chrisliu298
Send prompts to ChatGPT via Chrome browser automation and collect responses. Use this skill whenever the user wants to ask ChatGPT something, delegate a task to ChatGPT, get ChatGPT's perspective, or use ChatGPT's reasoning capabilities. Triggers on "ask chatgpt", "send to chatgpt", "use chatgpt", "chatgpt", "have chatgpt", or "chatgpt". Accepts two arguments: <model> <effort> where model is instant/thinking/pro and effort is low/medium/high/xhigh. Example: chatgpt thinking xhigh, chatgpt pro high, chatgpt instant.

0
installs
▌
Autoresearch · chrisliu298 bundle
Autonomous experiment loop faithful to Karpathy's autoresearch. Set up a branch, establish a baseline, then loop forever: edit, commit, run, measure, keep or revert. Use when asked to "run autoresearch", "optimize X in a loop", or "start experiments".

0
installs
▌
Add Paper · chrisliu298
Add a paper, technical report, or tool to the awesome-on-policy-distillation list (a curated collection of on-policy distillation (OPD) methods for LLMs). Use this skill whenever the user provides an arxiv URL, paper link, or asks to add a paper/report/tool to the list. Also trigger when the user pastes a URL and says something like "add this", "include this paper", or "put this in the list".

0
installs
▌
Citation Assistant · chrisliu298 bundle
Add verified citations to academic papers and proposals. Use when working on LaTeX documents that need citations, finding papers for a specific section, verifying existing citations, or when user mentions "cite", "citation", "references", "bibliography", or asks to find related work. Supports Semantic Scholar API, Exa MCP, and DOI-based BibTeX retrieval. Never hallucinate citations. Invoke with "citation-assistant" or "citation-assistant [section]".

0
installs
▌
Karpathy Guidelines · chrisliu298 bundle
Guidelines for effective LLM-assisted coding based on Andrej Karpathy's observations. Use this skill when writing code with an LLM to avoid common pitfalls: overcomplication, bloated abstractions, dead code, wrong assumptions, and 1000-line solutions that should be 100 lines. Applies to all coding tasks. Invoke with /karpathy-guidelines.

0
installs
▌
Last Call · chrisliu298 bundle
Session-end quality review. Deep-reads every changed file and evaluates correctness, documentation debt, architecture, and session completeness. Use when wrapping up a session, or when user says "final check", "anything I missed", "review before I'm done", or invokes "last-call". Do NOT use mid-session for incremental reviews; use /review-pr instead.

0
installs
▌
Nanorepl · chrisliu298 bundle
Build minimal reimplementations of complex projects following Karpathy's nano/micro/min philosophy. Analyzes a target project to identify its irreducible algorithmic core, decomposes it into progressive layers, and guides incremental building with verification at each step. Works for any domain. Triggers on "build a minimal version of", "nanorepl", "reimplement from scratch", "strip down to the core", "build a micro/nano/mini version", or when a project URL is provided with a request to understand or rebuild its core. Invoke with "nanorepl".

0
installs
▌
Chatgpt · chrisliu298
Send prompts to ChatGPT via chrome-devtools-mcp and collect responses. Use this skill whenever the user wants to ask ChatGPT something, delegate a task to ChatGPT, get ChatGPT's perspective, or use ChatGPT's reasoning capabilities. Triggers on "ask chatgpt", "send to chatgpt", "use chatgpt", "chatgpt", "have chatgpt", or $chatgpt. Accepts two arguments: <model> <effort> where model is instant/thinking/pro and effort is low/medium/high/xhigh. Example: $chatgpt thinking xhigh, $chatgpt pro high, $chatgpt instant.

0
installs
▌
Jina · chrisliu298
Fetch web pages or search the web via Jina AI (r.jina.ai / s.jina.ai). Use when WebFetch fails, returns empty/garbage, or for JS-heavy sites: x.com, Notion, SPAs, PDFs at a URL, Cloudflare-gated pages. Also use for grounded web search with full-page content. Triggers on "fetch this URL", "read this page", "search for", or after a WebFetch error. Do NOT use for simple static HTML.

0
installs
▌
Push · chrisliu298
Commit and push changes to the remote ("commit and push", "push this", "ship it", "save and push", "push"). The agent inspects the diff and picks single-commit or atomic mode automatically: atomic when changes span unrelated units (mixed categories, multiple distinct areas), single commit for one focused change. Do NOT create a pull request.

0
installs
▌
Todo · chrisliu298 bundle
Maintain a single TODO.md at the project root so task state survives session boundaries, compaction, and multi-day gaps. Use only for work large or complex enough that the native in-session checklist can't track it well — multi-day refactors, migrations, or anything spanning several sessions. Also trigger on "/todo", "track this", "checkpoint", or "save for next time". Do NOT trigger on small one-shot tasks (a single bug fix, rename, or one-liner), work the native checklist handles fine, in-session-only todo tracking, or when there's no clear project root.

0
installs
▌
Xurl · chrisliu298
Interact with X (Twitter) via xurl, the official X API CLI: post, reply, quote, delete, search, read timelines/mentions, like, repost, bookmark, follow, DM, upload media, raw v2 endpoints. Triggers on "post to X", "tweet this", "reply to [X post]", "search X for", any x.com/status URL, or "xurl". Do NOT trigger on bare "X" meaning the letter, a variable, or an unknown.

0
installs
▌
Crons · chrisliu298 bundle
Maintain a durable, version-controlled manifest of a project's recurring Claude Code cron jobs (the /loop + CronCreate fleet) and re-arm them after they are lost. Claude crons are session-only — durable:true is ignored, CronList truncates prompts, and recurring jobs auto-expire after 7 days — so the live state is unrecoverable; this skill keeps crons/*.cron as the source of truth, renders CRONS.md, and emits the exact CronCreate/CronDelete calls to re-arm. Use on session start, after a compaction, reboot, or 7-day expiry to check whether your crons are still armed and restore them; when you run many /loop crons in one project; or on "crons", "manage my crons", "re-arm crons", "are my crons still running", "cron manifest". It is a preparer, not an actuator: it never calls the Cron* tools and never claims a cron is "armed" (no false assurance) — you run CronList and execute the emitted calls. Do NOT use for a one-off cron (use CronCreate directly) or a cache heartbeat (use keep-warm).

0
installs
▌
Prism · chrisliu298 bundle
Dispatch multiple independent agents to answer the SAME complete question from different analytical lenses, then synthesize. Reach for this by default, without waiting to be asked, when redundant cross-model judgment could change the decision: ambiguous architecture/design tradeoffs, high-stakes or hard-to-reverse changes, competing root-cause hypotheses, or failure-mode-sensitive reviews — and whenever you would otherwise spawn 2+ independent reviewers for one question. Skip for trivial lookups, deterministic transforms, routine small edits, mechanical syncs, and single-correct-answer tasks. With no leading number, autonomously decide N and gpt-pro from the question every time (N=1 is the bottom-rung anchor, never a lazy default); with explicit numbers `prism N M` (e.g. `prism 2 1` = 2 of each model + 1 gpt-pro lens; M is a count, not a flag), honor them verbatim and skip auto-sizing. There is no reasoning-effort knob. Scale above the anchor only when decision risk justifies the 7N-agent cost.

0
installs
▌
Relay · chrisliu298 bundle
The ONLY way to call GPT (a.k.a. Codex), Grok, GLM, Kimi, DeepSeek, or MiMo. Use whenever the user wants to ask, delegate to, or get a second opinion from GPT, Grok, GLM, Kimi, DeepSeek, or MiMo. Do NOT run the codex, grok, glm, kimi (km), deepseek (ds), or mimo (mm) CLI directly — from the main agent or a subagent; always use this skill's relay call command. Triggers on "ask/have/send to/get/delegate to gpt/codex" or the same with "grok"/"glm"/"kimi"/"deepseek"/"mimo", "second opinion", "relay".

0
installs
▌
Deslop · chrisliu298
Remove AI-generated slop from code changes. Checks the diff against main and cleans up unnecessary comments, defensive checks, single-use variables, redundant casts, and style inconsistencies introduced by AI. Use when the diff looks bloated after a coding task, or when user says "clean up", "remove slop", "deslop", "simplify the diff", or invokes "deslop". Do NOT use for deliberate refactoring or feature changes.

0
installs
▌
Digest · chrisliu298
Re-layer the previous response (or a named file) into a fast-to-skim form: answer first, a review-map or key-findings short version, action-changing blockers only, detail pulled on demand. Restructures for fast reading WITHOUT dropping substance. Use when a reply came out dense, or the user says "tl;dr", "too long", "wall of text", "make this skimmable", "digest that", or invokes "digest". Do NOT use to summarize-away detail, on already-short replies, or to clean code (use deslop).

0
installs
▌
Recall · chrisliu298 bundle
Recall a detail from THIS project's PAST Claude sessions — a fact, value, name, decision, preference, or constraint the user now refers back to but that was never written into docs or code — by lexically searching all prior session transcripts for this cwd, loading the best match into your working context, and printing ONE confirmation line (date + session + gist) so a wrong match is caught before you act. Trigger on "/recall" or natural references to something stated in a prior session you're expected to remember: "what did we say about X", "the value we used for X", "like I said / as I mentioned earlier", "remember when we…", "what was my preference for…", "didn't we already settle X". Do NOT trigger on ordinary past-tense narration that carries its own content ("earlier I ran the tests"), git/docs/code lookup, broad "catch me up" summaries, curated cross-session facts (memory), or task state (todo). Treat any recalled item as evidence, not current truth — re-verify before acting.

0
installs
▌
Goal Loop · chrisliu298 bundle
Drive a goal through review→fix iteration by composing goal-elicit, goal-drive, and multi-model review (prism, Claude-only). A thin, modeless, stepped loop — one phase per invocation (elicit → spec-review → drive → review → you pick which fixes apply → fix → re-review) — with state on disk so it resumes after interruption or compaction. It auto-handles the ~83% of out-of-scope/unmapped findings (you never see them) and surfaces only the small actionable batch to confirm — never blind-applies (unsafe). Use for "loop this to done with review", "close the review-fix loop", "goal-loop". Review needs Claude (prism), degrades off-Claude. Interactive by default; **`--auto`** runs it unattended/headless under Claude Code's native `/goal` — both gates become fail-closed policies (oracle-gated safe-subset auto-fix + a deferred morning queue), never blind-apply. Does NOT interview (that's goal-elicit). Skip for one-off edits and lone reviews.

0
installs
▌
Keep Warm · chrisliu298
Schedule a recurring heartbeat that keeps Claude Code's 1-hour prompt cache from expiring during long unattended tasks. Use when starting a multi-hour run, or when the user says "keep cache warm", "warm cache", "stay warm", "heartbeat", "keep alive", or invokes "keep-warm". Pass "stop", "cancel", or "off" to cancel an active heartbeat.

0
installs
▌
Goal Drive · chrisliu298 bundle
Drive an existing goal artifact to verified done — a GOAL.md contract, a JSON checklist, or a phased design doc (produced by goal-elicit, or hand/script-written). Executes modelessly: process the next unit, verify acceptance, update state, optionally commit, repeat — stopping only on exceptions (blocked unit, scope drift, unclear authority, an irreversible action). Use for "drive this", "execute this plan/checklist", "process this in batches", "run the phases", "carry this to done". Does NOT interview — if the goal is unclear or no artifact exists, run goal-elicit first. Skip for one-off edits and factual questions.

0
installs
▌
Codex First · chrisliu298 bundle
Claude Code only. Delegate trusted, scoped implementation to `codex exec --yolo` — building from a spec, refactors, mechanical migrations, bug fixes with a repro, tests, CI/dependency/tooling, bulk code reading — while Claude writes the spec, picks `-e` effort, and reviews every diff. Auto-routes when enabled; also `/codex-first <task>`. Keep design, ambiguous specs, tiny edits, secrets/session-only tools, and all commit/push/GitHub ops in Claude. Unsandboxed: the review gates the diff, not the machine.

0
installs
▌
Exec Status · chrisliu298 bundle
Maintain one plain-English STATUS.md at the working-directory root — a living executive briefing a non-technical stakeholder grasps in ~30 seconds: what the run is trying to do, how it's going, what it's found, and whether to worry (no jargon, no logs). Use for work the user kicks off and walks away from, or that spans many steps or sessions unwatched: autonomous research/experiment runs, hyperparameter sweeps, long migrations, overnight or multi-day jobs, "go run this and keep me posted" — and on resume in a directory that already holds a STATUS.md for an active run. Also trigger on explicit asks: "status report", "executive summary", "CEO update", "I keep losing track of progress", or "exec-status" by name. Do NOT use for in-session work you're watching, single edits or bug fixes, quick Q&A, or work with no durable working directory. Zero-context-readable — complements todo (operator checklist), autoresearch (experiment state), and digest (reply reformat); does not replace them.

0
installs
▌
Goal Elicit · chrisliu298 bundle
Interview the user and write a verifiable goal artifact — a Goal Contract (GOAL.md), a JSON checklist, or a phased design doc, chosen by triage. Never executes the goal or invokes downstream tools — it writes the artifact and stops; hand off to goal-drive (or any agent) to execute. Use for "clarify what I want", "define done_when/acceptance criteria", "make this unambiguous", "plan this out", XY-problem requests, or untestable success. Skip trivial edits, factual questions, routine commands, and code review.

0
installs
▌
Interviewer · chrisliu298 bundle
Conducts mock technical interviews for AI/ML topics. Use when the user wants to practice for interviews, test understanding, or be quizzed on ML concepts. Triggers on "interview me", "quiz me", "mock interview", "practice questions", or "interview". Supports topic-based or material-based questioning with multiple difficulty levels.

0
installs
▌
Mental Seal · chrisliu298 bundle
Hold ONE supreme priority front-of-mind for a whole task or session — a single visible, user-owned vow in SEAL.md that resists distraction and scope-drift and persists on disk across context compaction. Use when the user says "mental seal", "seal this", "the one thing", "top priority", "non-negotiable", "hold this above all", "don't lose sight of this", or "/seal"; and when a project's SEAL.md is active (recall is anchor-driven, not an automatic session hook). Maintain exactly one active seal; recall it at decision points; surface conflicts instead of silently obeying; discharge only on verified completion, expiry, or explicit user resealing. NOT a task list (use todo), a stakeholder briefing (use exec-status), a recurring heartbeat (keep-warm), or a fact store (memory) — it is one always-present priority anchor with first-class enforcement and release semantics.

0
installs
▌
Arxiv Reader · chrisliu298
Read arxiv papers via their TeX source for full-fidelity math, tables, and figures. Use whenever the user shares an arxiv URL or paper ID and wants to discuss or understand the paper — even without "read". Triggers on arxiv URLs (abs, pdf, html), bare IDs like "2401.12345", or "explain this paper". Do NOT use for note generation (use note-gen).

0
installs
▌
Skill Wiring · chrisliu298
How this dotfiles repo wires skills, MCP servers, and Claude plugins — where to edit each kind of skill, the SKILLS/MCP_SERVERS/PLUGINS tables in dotfiles.sh, manual-skill toggling, and the authoring rules for universal (Claude/Codex/Grok/Pi) skills. Use when adding, editing, moving, scoping, enabling, or disabling a skill, MCP server, or plugin in this repo, or when a skill's symlink or agent targeting looks wrong.

0
installs
▌
Dual Subagent · chrisliu298 bundle
Delegate read-only tasks to Claude Code and Cursor Agent concurrently, including research, reading, analysis, diagnosis, checks, review, and audit. Use when the user requests this skill, asks both agents, or the applicable delegation rules route a read-only task here. Tasks requiring file or external-state changes belong to native subagents.

0
installs
▌
Gpt Pro Relay · chrisliu298 bundle
Send a prompt to ChatGPT Pro via gpt-pro-relay on macmini — over SSH from any other machine, or directly when invoked on macmini itself. Use for "ask gpt-pro", "send to gpt-pro", "use gpt-pro", "Pro take", "ask the deep model", or "second opinion from chatgpt pro". The wrapper (not the caller) polls through flaky networks; the agent fires one backgrounded call and waits for the notification.

0
installs
▌
Skill Cleaner · chrisliu298 bundle
Audit Codex/OpenClaw skills: loaded roots, duplicate skills, unused skills, prompt-budget costs, compact descriptions.

0
installs
▌
Claude Subagent · chrisliu298 bundle
Obtain an independent read-only review from Claude Code. Use when the user asks to "ask Claude", "use Claude as a reviewer", get a "Claude take", or obtain an independent Claude review or second opinion. Do not use this skill to implement, complete, advance, or operate a task; use Codex's native subagents when delegated execution is requested.

0
installs
▌
Cursor Subagent · chrisliu298 bundle
Obtain an independent read-only review from Cursor Agent using a pinned GPT-5.6 Sol model. Use when the user asks to "ask Cursor", "use Cursor as a reviewer", get a "Cursor take", or obtain an independent Cursor review or second opinion. Do not use this skill to implement, complete, advance, or operate a task; use Codex's native subagents when delegated execution is requested.

0
installs
▌
Prompt Engineer · chrisliu298 bundle
Guide for writing or refining prompts for Claude or GPT-5.6 Sol, distilled from each vendor's official best practices. Use for new prompts, debugging an existing prompt, or "prompt engineering", "system prompt", "Claude/GPT/OpenAI prompt". Accepts arg: "claude" or "gpt".

0
installs

34 published skills