vLLM: high-throughput LLM serving, OpenAI API, quantization.
Langue du texte source : anglais
Menu
Skills dans ce dépôt
SkillsMP a collecté 447 skills depuis jleechanorg/claude-commands. Ouvrez un skill pour examiner sa source et ses détails.
jleechanorg/claude-commandsAffichage de 40 skills collectés sur 447.
vLLM: high-throughput LLM serving, OpenAI API, quantization.
Langue du texte source : anglais
vLLM: high-throughput LLM serving, OpenAI API, quantization.
Langue du texte source : anglais
SAM: zero-shot image segmentation via points, boxes, masks.
Langue du texte source : anglais
Expert guidance for GRPO/RL fine-tuning with TRL for reasoning and ta[REDACTED_OPENAI_KEY] model training
Langue du texte source : anglais
Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods. Use when fine-tuning large models (7B-70B) with limited GPU memory, when you need to train <1% of parameters with minimal accuracy loss, or for multi-adapter serving. HuggingFace's…
Langue du texte source : anglais
Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. Use when need RLHF, align model with preferences, or train from human feedback. Works…
Langue du texte source : anglais
Expert guidance for fast fine-tuning with Unsloth - 2-5x faster training, 50-80% less memory, LoRA/QLoRA optimization
Langue du texte source : anglais
TEST
Langue du texte source : Indéterminée
Recipe for cleaning up an already-polluted PR — one whose branch was created from the wrong base (or merged another PR's full history) and now carries unrelated commits. Produces a fresh branch from origin/main, cherry-picks only the load-bearing diff, and…
Langue du texte source : anglais
Create, read, edit Word .docx documents and templates.
Langue du texte source : anglais
Use already-installed `gws` or `gog` (gogcli) credentials when Hermes's own Google OAuth setup has not been run. Reuses the refresh token from the working CLI to call Gmail/Calendar/Drive APIs directly via curl, avoiding a full OAuth re-setup. Activates when…
Langue du texte source : anglais
Generate the morning life-channel Slack digest — fetch unread Gmail, query next-24h Calendar across all known accounts, filter for "important" items, post to
Langue du texte source : anglais
Create, merge, split, fill, and secure PDF files.
Langue du texte source : anglais
Coordinate the user's annual US personal income tax return — gather W-2, 1099, 1095, brokerage crypto and foreign-account data, fill the Preparer's organizer questions, and queue the upload bundle for the Preparer's secure portal (TaxDome / Drivetime /…
Langue du texte source : anglais
Create, read, edit .pptx decks, slides, notes, templates.
Langue du texte source : anglais
Create, read, edit Excel .xlsx spreadsheets and CSVs.
Langue du texte source : anglais
Diagnose and fix Python 3.14 `DaemonThreadPoolExecutor` / `ThreadPoolExecutor` AttributeError on `_initializer` or `_initargs`. Triggers on the literal error string `'DaemonThreadPoolExecutor' object has no attribute '_initializer'` (or `_initargs`), or any…
Langue du texte source : anglais
Read grok.com/share/* via Aside REPL openTab.
Langue du texte source : anglais
Thin pointer — canonical /repro workflow lives in WorldArchitect repo at .claude/skills/repro-twin-clone-evidence/SKILL.md. File gh issue, create draft PR, copy campaign, reproduce, verdict.
Langue du texte source : anglais
Write publication-ready ML/AI papers for NeurIPS, ICML, ICLR, ACL, AAAI, COLM. Use when drafting papers from research repos, structuring arguments, verifying citations, or preparing camera-ready submissions. Includes LaTeX templates, reviewer guidelines, and…
Langue du texte source : anglais
Write ML papers for NeurIPS/ICML/ICLR: design→submit.
Langue du texte source : anglais
Investigate a question against high-trust primary sources and capture the findings as a Markdown file in the repo. Use when the user wants a topic researched, docs or API facts gathered, or reading legwork delegated to a background agent.
Langue du texte source : anglais
/roadmap slash command + automated Slack-thread-audit skill. Runs a 48h Slack thread sweep across user channels, classifies each thread as finished/pending/needs-human-decision, runs /nextsteps on each pending thread, drafts auto-action proposals, pushes the…
Langue du texte source : anglais
Audit ALL home/~/backup/sync/snapshot-related jobs + scripts on a Mac for any that push personal home-dir content to PUBLIC GitHub repos. Trigger when a backup/ leak is found in a public repo, when the user says "double check all the backup jobs", or when any…
Langue du texte source : anglais
Prevent AI agents from publishing live credentials in outbound GitHub/Slack/email/gist artifacts. Trigger when an agent reads git remote -v, .git/config, gh auth status -t, env, printenv, cat ~/.bashrc, an AO runner yaml, the ao-go-daemon plist, or any…
Langue du texte source : anglais
After a complex / multi-turn session (>5 tool calls, multiple spawns, multiple commits, or user corrections), use this skill to extract /learn artifacts (memory + skills) BEFORE ending the session. Triggers on "/learn", "/learn from this session", "anything…
Langue du texte source : anglais
Stop replying "Idle" indefinitely when the user is silent — recognize three wrapper-side re-injection patterns (echo-of-self, queued-tick, system-marker) and exit the loop cleanly. Triggers when the assistant is about to reply "Idle" / "no new instruction" /…
Langue du texte source : anglais
Turn any feature, script, or workflow into a properly-skilled, tested, auditable Hermes skill. Runs the gbrain-derived 11-item Skillify Completeness Contract against the target and creates all missing artifacts.
Langue du texte source : anglais
Detect and stop responding to echo loops in Slack threads. When the gateway or babysit cron repeatedly re-injects prior assistant turns as if they were new user messages, replying with anything more than a single-word acknowledgment burns iteration budget…
Langue du texte source : anglais
Draft social-media posts for LinkedIn, Hacker News, Twitter/X, Reddit, Threads, Facebook, Instagram, Mastodon, Bluesky, and Dev.to. Each draft is staged in an Aside browser tab (signed-in persistent session), screenshotted, and surfaced for human review.…
Langue du texte source : anglais
When the user asks to compare, evaluate, or compose AI coding agent methodologies and skills frameworks (Superpowers, GSD Core, grill-me, OpenSpec, BMAD, Spec-Kit, Kiro, gstack, GSTACK, custom /skill packs). Use for "should we install X?", "how do X and Y…
Langue du texte source : anglais
Run grep before citing specs.
Langue du texte source : anglais
Verify that a CLI tool or daemon honors a specific environment variable before claiming the var 'works'. Combines four signals in priority order: (1) binary-string grep on the compiled executable as ground truth, (2) official vendor docs, (3) live process…
Langue du texte source : anglais
Guidelines for performing thorough code reviews with security and quality focus
Langue du texte source : anglais
Triage when a single campaign (or single LLM/prompt surface) has accumulated multiple sibling bug investigations — the prompt-discipline family where "LLM emits right structured field, render/persistence/propagation path drops it", the now-merged NPC-status…
Langue du texte source : anglais
Diagnose and fix module-global boolean/scalar flags used as schema or feature gates that race across concurrent writers — specifically the pattern where a single in-process bool gates "have we confirmed the target table has these columns / feature is…
Langue du texte source : anglais
Safe multi-patch editing in a single session — avoid the partial-read self-revert trap. Use when editing a file with multiple sequential `patch` tool calls, when a patch warning says 'file was last read with offset/limit pagination', when reviewing your own…
Langue du texte source : anglais
Write a markdown plan to .hermes/plans/; no execution.
Langue du texte source : anglais
Compact an LLM prompt (markdown file in `**/prompts/**` or equivalent) by 10-50% without breaking the test-anchored contract that pins its content. Trigger when the user asks to "compact", "shrink", "trim", "shorten", "tighten", or "compress" a prompt file,…
Langue du texte source : anglais
Use when designing, refactoring, or auditing Claude/Codex slash commands under `~/.claude/commands/*.md` or repo-local `.claude/commands/*.md`. Covers the project-agnostic (user-scope) vs repo-specific (repo-local) split, the bidirectional-pointer contract,…
Langue du texte source : anglais