AI & ML

nikhil8bph Skill

Acts as a Solution Architect who requests/requires a Business Requirement Document (BRD) and produces a production-grade Application Development Planning Document covering feasibility, system architecture and topology, database schemas, security architecture, resiliency patterns, and integration strategy - intentionally excluding detailed API contracts. Use when a BRD exists (or must be requested from the user) and the task is to decide the overall tech stack, architecture pattern, and data model before any API endpoint or contract design begins.

alephic-ai Skill

A reference skill that exercises every supported frontmatter field — the six base Agent Skills fields plus all thirteen Claude Code extensions. Use when testing skref's --allow-claude-fields support or as a worked example of the full Claude frontmatter surface.

colbytimm Bundle

Author or improve an AGENTS.md file — the open-standard Markdown config that guides AI coding agents (Codex, Copilot, Cursor, Claude Code, Aider, etc.) in a repo. Use whenever the user wants to create, write, review, trim, or fix an AGENTS.md (or CLAUDE.md/AGENT.md), set up agent instructions for a project or monorepo package, or asks why their existing agent file isn't helping. The discipline is token efficiency: every line must earn its place, because a bloated file measurably degrades agent output — worse than no file at all.

colbytimm Bundle

Install and configure a Powerlevel10k-styled status line for terminal coding agents that support a command-backed statusline (Claude Code and GitHub Copilot CLI). Ships a portable, dependency-light Node renderer (statusline.js) that reads the host agent's session JSON on stdin and prints a powerline bar — agent + model, git branch, context-window usage, a fill gauge, elapsed time, and line changes — plus a legacy zsh variant. Use when the user wants to set up, theme, fix, or port a coding-agent statusline, asks why their Claude/Copilot status bar is blank or stuck on "waiting for first exchange", or wants one statusline that works across macOS, Linux, and Windows. Codex CLI and Gemini CLI render their status lines internally and cannot be driven by this skill.

collectifweb Bundle

Filtre d'écriture qui détecte et corrige les tics LLM dans les textes français pour les rendre humains. Déclencher ce skill dès que l'utilisateur tape /humanize, demande d'« humaniser » un texte, de « nettoyer le style IA », de « virer le slop », de « rendre naturel », ou prépare un contenu destiné à être envoyé à quelqu'un (rapport, mail client, message, article, post). Aussi quand il mentionne les tirets cadratins, les majuscules après deux-points, le style robot, le ton IA, ou dit que « ça sonne ChatGPT ». Utiliser ce skill AVANT d'envoyer tout livrable : rapport, e-mail client, message, article, post LinkedIn, document, présentation textuelle. Si le texte sort de Claude et va vers un humain, ce skill s'applique.

devonjones Skill

Compact the current conversation into a handoff document so a fresh agent (new session, different machine, or a teammate's chat) can pick up the in-flight work without ramping from zero. Use when: (1) ending a long session with state that won't survive compaction or memory alone, (2) handing work to a teammate or a different agent (Claude Code → Codex, Claude → Cursor), (3) parking work for later when the to-resume context is too large to reconstruct from git history + beads alone.

devonjones Skill

Validate the agent reviewers that gate a project's quality (the pr-review-loop roster). Mine every PR's review threads + the dream-marker stream into per-reviewer scorecards (acceptance rate, false-positive/won't-fix/reopen rates, taste overrides), then propose roster changes (keep / refine / retire). Use when: (1) the user asks to validate/audit/tune their PR reviewers or AGENT-REVIEWERS.md, (2) deciding whether a reviewer earns its keep or should be retired/refined, (3) a periodic reviewer-health pass. For mining general session lessons, use the `dream` skill instead.

devonjones Bundle

Generate, edit, and compose images using Google's Gemini 3 Pro Image model (Nano Banana Pro). Use this skill when the user asks to create images, generate visuals, edit photos, compose multiple images, create logos, thumbnails, infographics, product shots, or any image generation task. Supports text-to-image, image editing, multi-image composition (up to 14 images), iterative refinement, aspect ratio control, and Google Search-grounded image generation for real-time data visualization.

devonjones Bundle

Manage the PR review feedback loop: monitor CI checks, fetch review comments, and iterate on fixes. Use when: (1) pushing changes to a PR and waiting for CI/reviews, (2) user says "new reviews available", (3) iterating on PR feedback from Gemini, Cursor, Claude, or other reviewers, (4) monitoring PR status. Supports multiple review bots: Gemini Code Assist, Cursor Bugbot, and Claude agent fallback. Also supports custom agent reviewers defined in AGENT-REVIEWERS.md for focused reviews (security, DRY, etc.). Automatically detects priority levels from different bot formats and handles rate limits.

devonjones Bundle

Download a YouTube video and extract frames at specified timestamps with perceptual hashes; also discovers candidate timestamps via ffmpeg scene-detect + per-second pHash run-grouping; analyses optical flow over a sub-range and composes a sprite strip showing a motion arc. Use when capturing specific moments from a video as PNG images, or when a downstream skill (e.g. youtube-synthesizer) needs a high-recall list of where things change in the video. Mechanical only — no LLM calls.

fast-agent-ai Bundle

LSP-enable a Python, TypeScript, or Rust repository for fast-agent development. Use when creating or refreshing agent cards with LSP function tools, configuring ty, typescript-language-server, or rust-analyzer integration, or scoping repo-local code navigation safely.

fast-agent-ai Skill

Build, review, or refactor GEPA optimization loops that use fast-agent. Use when Codex needs to design GEPA evaluators, Actionable Side Information, Pareto/frontier metrics, FastAgentReflectionLM, FastAgentBatchEvaluator, BatchRunner, EvalRun/CandidateRun artifact layouts, Trackio monitoring, or refactor OpenClaw/Birch-style prompt, policy, skill, batch, or artifact-generation optimizers.

fast-agent-ai Bundle

Guide for implementing fast-agent hooks and user-invoked command actions. Use when adding hook functions to agent cards or Python code for tasks like history compaction, saving sessions, modifying tool calls, managing agent lifecycle, or adding explicit slash-command add-ins with PluginCommandActionContext.

fast-agent-ai Skill

Investigate fast-agent session and history files to diagnose issues. Use when a session ended unexpectedly, when debugging tool loops, when correlating sub-agent traces with main sessions, or when analyzing conversation flow and timing. Covers session.json metadata, history JSON format, message structure, tool call/result correlation, and common failure patterns.

fast-agent-ai Bundle

Implement hook-based history compaction in fast-agent (rolling window, context threshold truncation, tool-result clearing, summary prompt compaction). Use when adding or extending compaction hooks and wiring them into agent cards.

immamdouhaboammar Skill

Use when a repository has a valuable agentic workflow, prompt chain, playbook, command, or runbook that should become a portable ChatGPT/Codex Skill without losing its real decision logic.

jinjie0703 Bundle

AI Agent 技能架构师与脚手架工具。指导设计、自动化生成符合双轨制规范（SKILL.md + README.md）的 Agent 技能骨架，并执行工业级安全与规范质检。在需要创建新技能、编写 Skill 规范、或对已有技能进行合规审计时使用。

jinjie0703 Bundle

企业级 Prompt 优化、结构化重构与对抗评测技能。将模糊的自然语言需求升级为高召回、带防越狱护栏 (Guardrails) 与精确 JSON Schema 约束的工业级 Prompt，并自动生成对抗压力测试用例。在需要编写或优化 Prompt、设计系统提示词、规范大模型输出格式或进行防注入测试时使用。

joheredi Skill

Produces a grounded implementation design by comparing viable approaches, recommending one, running parallel multi-model critique, reconciling feedback, and revising the design.

jfcrenshaw Bundle

Guide to Firecrown, the DESC likelihood framework for cosmological parameter inference. Covers the Likelihood/Statistic/Systematic/ModelingTools architecture, two-point statistics (angular power spectra and correlation functions), systematic effects (intrinsic alignment, photo-z shifts, multiplicative shear bias, galaxy bias), sacc as the data container interface, and connectors to CosmoSIS, NumCosmo, and Cobaya. Also covers the TXPipe→Firecrown data flow via sacc files. Use this skill whenever the user imports firecrown, works with likelihood objects, builds a Firecrown analysis, connects to CosmoSIS or NumCosmo, reads or writes sacc files for inference, asks about DESC likelihood inference, or calls `load_likelihood` / `build_likelihood`.

jonny-1812 Skill

Create a brand-new Corebee account from the terminal — no dashboard, no browser login. Use when the user does NOT have a Corebee account yet and wants to set up AI customer support, start the free trial, get a chat/support widget, or asks to "set up Corebee for my product/site/company". One email approval, then you receive a scoped API key, their org, and a ready-to-paste widget snippet. If the Corebee MCP server asks you to authenticate and the user has no account, run THIS instead of sending them to a browser login.

jonny-1812 Skill

Triage and prioritize open support conversations by reviewing unassigned tickets, identifying urgent issues, and suggesting agent assignments based on workload. Use when the user asks to review their inbox, triage conversations, check what needs attention, handle unassigned tickets, balance workload, or see what is urgent.

jonny-1812 Skill

Analyze team capacity and workload distribution — see who is overloaded, who has bandwidth, and rebalance conversation assignments across the team. Use when the user asks about team capacity, workload, agent availability, or wants to redistribute work.

jonny-1812 Skill

Handle urgent customer conversations end-to-end — identify critical issues, assign to the best available agent, log escalation notes, and optionally draft an immediate response. Use when the user says something is urgent, needs escalation, or wants to handle a critical ticket.

lyupro Skill

Tighten and de-bloat an LLM prompt while preserving intent.

lyupro Bundle

Summarise the authorship and recent change history of a file via git blame + LLM polish.

lyupro Bundle

Generate a CHANGELOG.md entry from git log since the last tag, polished by the LLM.

poudatmorteza Skill

Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.

benjaminshoemaker Skill

Search past conversations with Claude Code and Codex CLI. Finds matching excerpts across all sessions with metadata (date, project, agent). Use when looking for a past discussion, decision, or piece of context from any previous session.

benjaminshoemaker Bundle

Use when determining how an AI coding agent can autonomously run and verify a specific user flow end to end before implementation.

jxnl Bundle

Audit, de-slop, parameterize, modularize, or safely clean up AI-generated or AI-shaped backend/general code. Use for Python, TypeScript, or other implementation diffs that may contain duplicate helpers, fixture hacks, hard-coded test data, over-defensive control flow, broad exception wrappers, config-bag or boolean-mode soup, speculative scaffolding, hallucinated APIs/dependencies, local-idiom drift, brittle tests, or maintainability/safety/performance gaps after a feature, bugfix, prototype, or agent pass.

jxnl Bundle

Audit pasted chatbot output, AI-cleanup diffs, wiki drafts, Markdown/MDX/docs, and source-backed articles for generic AI fluff, LLM writing tells, weak audience model, lack of theory of mind, inflated significance, vague attribution, leaked tokens, placeholders, broken markup, fabricated or mismatched citations, and detector false positives. Use for requests like "AI writing audit", "check for AI slop", "find writing fluff", "does this sound like ChatGPT", "cleanup LLM tells", "make this less generic", "verify these citations", "detector flagged this", or text containing turn0search0, oaicite, oai_citation, contentReference, utm_source=chatgpt.com, malformed references, or wrong target-format markup.

jxnl Bundle

Help address review/issue comments on the open GitHub PR for the current branch using `oai_gh` or `gh`; verify auth first and prompt the user to authenticate if not logged in.

pantheon-org Bundle

Evaluate, score, and remediate agent skill collections using a 9-dimension quality framework (Knowledge Delta, Mindset, Anti-Patterns, Specification Compliance, Progressive Disclosure, Freedom Calibration, Pattern Recognition, Practical Usability, Eval Validation). Performs duplication detection, generates remediation plans with T-shirt sizing, enforces CI quality gates, validates artifact conventions, tracks score trends, and provides a native eval runner with LLM-as-judge (BYOK, provider-agnostic). Use when evaluating skill quality, auditing SKILL.md files, scoring agent skills, generating remediation plans, detecting duplicate skills, validating skill format, enforcing quality gates, optimizing for A-grade publication, comparing audit baselines, batch skill assessments, or checking skill registry compliance. Triggers: 'check my skills', 'skill audit', 'improve my SKILL.md', 'quality check', 'A-grade scoring', 'quality gates', 'eval validation', 'audit all skills', 'remediation plan', 'skill judge', 'dimens

pantheon-org Bundle

Review .context/plans/*.md files using 3 independent subagent reviewers: Technical (feasibility, gaps, soundness), Strategic (scope, alignment, priority), and Risk (blind spots, edge cases, failure modes). The skill FIRST asks the user which models to assign to each reviewer — presenting available options based on the user's environment (OpenCode Zen, Go, Claude Code, or BYOK). Each reviewer receives the same plan with a role-specific lens and returns structured feedback. The main agent consolidates all three perspectives. Triggers: 'review this plan', 'audit plan', 'plan review', 'check my plan', 'what's wrong with this plan', 'review all plans', 'plan quality check'. Do NOT use for plans not in .context/plans/, one-off notes, or external documents.

pantheon-org Bundle

Manage agent behavioural rules in `.agents/RULES.md`. Use when the user says 'new rule', 'add a rule', 'record that', 'make a rule that', or asks to codify an agent directive. Load `.agents/RULES.md` to check existing rules before adding a new one. DO NOT use for ephemeral notes, one-off instructions, or instructions that belong in AGENTS.md itself. Triggers: 'new rule', 'add a rule', 'record that', 'make a rule', 'codify this', 'rule about', 'create a directive'.