summaries of all the papers I read

Research the last 30 days across Reddit, X, YouTube, Hacker News, Polymarket, GitHub, and the web, then synthesize a grounded brief for an AI agent.

Turn any technical book PDF into a Claude Code skill — ready to study, reference, and use while you work.

A cross-agent research paper toolkit that transforms papers into learning environments with summaries, code demos, and a local web viewer for Claude Code, Codex, OpenCode, and DeepSeek Harness.

Audio generation skill — jingles, beds, voiceover, and sound effects. Routes music requests to Suno V5 / Udio / Lyria, speech to MiniMax TTS / FishAudio / ElevenLabs V3, and SFX to ElevenLabs SFX or AudioCraft. Output is one MP3/WAV file saved to the project folder.

Structured medical case presentation for clinical rounds, conferences, and documentation. Generates SOAP-format or narrative case reports with physiologically accurate vitals, labs, and evidence-based plans. Use when the brief mentions "case report", "case presentation", "SOAP note", "clinical case", "ward rounds", "case summary", or "patient presentation".

收集、脱敏、预览并提交 Cherry Studio BUG、UI/UX 或功能反馈，默认提交到飞书。可在用户同意后调用内置诊断工具整理环境、错误日志、截图和用户导出的 trace，自动提交飞书表单或生成匿名上传 ZIP；也可安全解析反馈 ZIP 为表单字段。用户说“提交问题”“提交反馈”“上报 bug”“收集/上传错误信息”“整理日志/trace”“生成反馈包”，或描述 Cherry Studio 问题并希望记录时触发。只有明确要求 GitHub Issue 时才改用 issue-reporter。

Create a new skill in the current repository. Use when the user wants to create/add a new skill, or mentions creating a skill from scratch. This skill follows the workflow defined in .agents/skills/README.md and helps scaffold, validate, and sync new skills.

Prepare a new release by collecting commits, generating bilingual release notes, updating version files, and creating a release branch with PR. Use when asked to prepare/create a release, bump version, or run `/prepare-release`.

Autonomously improve a real artifact (code, training recipe, agent harness, data pipeline, prompt) against an objective and an evaluator, using Hypothesis Tree Refinement (HTR) from the Arbor paper. Use this whenever someone wants to iteratively optimize something over many experiments without overfitting — e.g. "get my model's eval score up", "improve this agent/harness", "tune this pipeline", "beat the baseline on this benchmark", "run a search over approaches and keep the best", "do an MLE-bench / Kaggle-style optimization", or any long-horizon "make this artifact better and don't just memorize the dev set" task. Trigger it even when the user doesn't say "Arbor" or "hypothesis tree" but describes repeated experiment-and-evaluate loops, branching exploration of competing ideas, or worries about a dev/test gap. Runs Claude itself as the coordinator with subagent executors in isolated git worktrees; for the standalone `arbor` CLI tool see references/arbor-upstream.md.

Observe the user's screen via screenpipe, detect repeated research workflows, match them against existing scientific-agent-skills, and draft new skills (or composition recipes that chain existing ones) for the patterns not yet covered. Use when the user asks to analyze their recent work and propose skills based on what they actually do. Requires the screenpipe daemon (https://github.com/screenpipe/screenpipe) running locally on port 3030 — the skill has no other data source and will refuse to run if screenpipe is unreachable. All detection runs locally; only redacted cluster summaries reach the LLM.

Adversarial code review that breaks the self-review monoculture. Use when you want a genuinely critical review of recent changes, before merging a PR, or when you suspect Claude is being too agreeable about code quality. Forces perspective shifts through hostile reviewer personas that catch blind spots the author's mental model shares with the reviewer.