Delegate heavy code generation to Codex CLI while Claude orchestrates, reviews, and surgically fixes. Use for any task involving 50+ lines of new code, boilerplate generation, test writing, or large refactors. Claude acts as lead engineer; Codex acts as developer. Multi-LLM aware — checks .claude/llm-mode.json before delegating and falls back to direct write if Codex is unavailable.
EnglishAutonomously crawl an entire website starting from a URL, following links to discover and scrape…
EnglishIndependently evaluate an implementation plan posted on a GitHub issue. Posts findings and adds plan-reviewed label.
EnglishScan LLM provider APIs for available models, fetch official docs for specs (parameters, context windows, capabilities, pricing), save timestamped doc snapshots, and optionally update a project's CLAUDE.md model table. Use when the user asks about current models or wants to refresh the model list — "scan models", "refresh models", "what models are available", "update model table", "model pricing", "check model IDs".
EnglishAI pair programming with Claude (builder) and a second model (advisor). The human observes and can intervene. Three styles: standard (Claude builds → advisor reviews), ping-pong (TDD-style alternation), and strong-style (advisor proposes → Claude implements). Use when asked for "pair session", "claude codex pair", "ai pair programming", or "/pair-session".
EnglishCreate, clone, read, run, and delete Salesforce Reports from the command line via the Analytics REST API, driven by the Salesforce CLI (sf). Org-agnostic — works against any org you've authed with sf. Includes a --setup flow (installs the sf CLI if missing, saves a default org/API version, and an optional Global Company filter for GAM roles). Use when the user says "create a Salesforce report", "clone a report", "generate a report via sf CLI", "list/run/delete reports", "report API", or "Analytics REST reports".
EnglishClean up, update, and upgrade this machine's package managers — Homebrew (formulae + casks), npm globals, and (when present) pipx and uv tools — safely and in parallel. Surveys what's outdated first, runs the safe upgrades, guards against known traps (a tool losing its CLI entrypoint on upgrade, a GUI app's binary shadowing a same-named CLI), and HOLDS session-critical cask updates (the terminal running the session, the VPN) for explicit confirmation. Use when the user says "update packages", "upgrade the machine", "clean up homebrew/npm", "update everything", "package cleanup", or "/update-machine".
English
AGENT SKILLS ATLASIncludes Codex Write, Evaluate Plan, Model Scan, and other workflows.
Claude Code
Codex
Cursor
GitHub Copilot
Gemini CLISupported by the Atlas install prompt. Skill-specific requirements vary.
Pack stats
What’s included
Codex Write
Delegate heavy code generation to Codex CLI while Claude orchestrates, reviews, and surgically fixes. Use for any task involving 50+ lines of new code, boilerplate generation, test writing, or large refactors. Claude acts as lead engineer; Codex acts as developer. Multi-LLM aware — checks .claude/llm-mode.json before delegating and falls back to direct write if Codex is unavailable.
Crawl
Autonomously crawl an entire website starting from a URL, following links to discover and scrape pages
Evaluate Plan
Independently evaluate an implementation plan posted on a GitHub issue. Posts findings and adds plan-reviewed label. Usage: /evaluate-plan
Model Scan
Scan LLM provider APIs for available models, fetch official docs for specs (parameters, context windows, capabilities, pricing), save timestamped doc snapshots, and optionally update a project's CLAUDE.md model table. Use when the user asks about current models or wants to refresh the model list — "scan models", "refresh models", "what models are available", "update model table", "model pricing", "check model IDs".
Pair Session
AI pair programming with Claude (builder) and a second model (advisor). The human observes and can intervene. Three styles: standard (Claude builds → advisor reviews), ping-pong (TDD-style alternation), and strong-style (advisor proposes → Claude implements). Use when asked for "pair session", "claude codex pair", "ai pair programming", or "/pair-session".
Salesforce Reports
Create, clone, read, run, and delete Salesforce Reports from the command line via the Analytics REST API, driven by the Salesforce CLI (sf). Org-agnostic — works against any org you've authed with sf. Includes a --setup flow (installs the sf CLI if missing, saves a default org/API version, and an optional Global Company filter for GAM roles). Use when the user says "create a Salesforce report", "clone a report", "generate a report via sf CLI", "list/run/delete reports", "report API", or "Analytics REST reports".
Update Machine
Clean up, update, and upgrade this machine's package managers — Homebrew (formulae + casks), npm globals, and (when present) pipx and uv tools — safely and in parallel. Surveys what's outdated first, runs the safe upgrades, guards against known traps (a tool losing its CLI entrypoint on upgrade, a GUI app's binary shadowing a same-named CLI), and HOLDS session-critical cask updates (the terminal running the session, the VPN) for explicit confirmation. Use when the user says "update packages", "upgrade the machine", "clean up homebrew/npm", "update everything", "package cleanup", or "/update-machine".
Quick start
- Choose your skills. Select individual skills above, or choose the entire pack.
- Copy the install prompt. Use Copy install prompt above the list.
- Send it to your agent. Paste into your coding agent’s chat. Your agent reviews the source and guides the setup.
Try it with your agent
After installation, start with a request like:
Use the Codex Write skill to help with my task.
Ask me for the context you need before getting started.Requirements & license
- A coding agent that can access your project files and install skills.
- Your agent checks each skill’s instructions, dependencies, and permissions during setup.
- Reported license for the skills shown: MIT. The author’s terms apply.