В каталоге 183 730 навыков в 14 категориях.
Launch a meta-judge then a judge sub-agent to evaluate results produced in the current conversation
Evaluate solutions through multi-round debate between independent judges until consensus
Launch an intelligent sub-agent with automatic model selection based on task complexity, specialized agent matching, Zer...
Execute tasks through systematic exploration, pruning, and expansion using Tree of Thoughts methodology with meta-judge...
Bifurcation creating/destroying equilibrium pair
Train sparse autoencoders to interpret model features.
Provides guidance for training and analyzing Sparse Autoencoders (SAEs) using SAELens to decompose neural network activa...
Make iOS 26+ Safari tint its status bar and floating tab bar to match the page, so a site looks native in Safari instead...
Get in the jeep and drive across the savanna, stopping at each item in a collection to observe, document, and design imp...
Use when scanning code for Safari/WebKit pitfall patterns and auto-fixing them — code that works in Chrome but breaks in...
Draft and fill Y Combinator SAFE templates — valuation cap, discount, MFN, pro rata side letter. Standard startup fundra...
Draft and fill Y Combinator SAFE contract templates — valuation cap, discount, MFN, pro rata side letter. Standard start...
Use when a request asks for a restart, rollback, production mutation, or the safest remediation process.
Agent with comprehensive Gemini safety settings configured for all categories
Agent with minimal Gemini safety settings using default threshold only
Agent with selective Gemini safety settings for specific categories
Sort a change proposal into safe mechanical items and items needing human judgment. Do NOT use to apply the changes.
Preview an `apt install` transaction with `--simulate` and flag dangerous removals (kernel, NVIDIA, mesa, DKMS, desktop...
Detects unsafe safe-area overrides and hardcoded inset values that hide content behind the Dynamic Island, notch, or hom...
Complete guide to handling safe areas in Capacitor apps for iPhone notch, Dynamic Island, home indicator, and Android cu...
Safety rules for files, metadata and tools in any Editor job. Use alongside every other Editor skill, and whenever a fil...
Apply one logical code change with mandatory impact mapping, git snapshot, verification, and automatic revert on failure...
The protocol for changing anything that can move real money — order placement, closing, position sizing, the MT5 bridge,...
Turn a confirmed specification, root-cause finding, or impact analysis into an atomic, verifiable, reversible implementa...
Reclaim disk space safely on Windows, macOS or Linux with a hash-approved deletion plan; the packaged helper deletes onl...
Ensures code modifications follow safe practices - never remove imports/items without verifying usage, check if truly ob...
Review a concrete diff or PR through three independent axes: repository conformance, change-intent fidelity, and operati...
コマンド実行の安全ポリシーを適用し、破壊的・外部影響のある操作は事前確認する
git commit を行う際に自動発動。禁止ディレクトリ（ai/, .serena/）のコミット防止と、論理的なコミット分割を支援する。
⚠️ MANDATORY - YOU MUST invoke this skill when committing. Complete commit workflow with all safety checks. Invokes secu...
Ship local changes to origin/dev with hard-enforced git safety policy, integrator-shell lock/guard flow, validate-change...
Pre-ship safety proof for replacing a live implementation with a canonical one. Use when repointing nodes, tools, or ser...
Rigor Debug / Rigor Audit skill for deep learning research work. Use when the user pastes a traceback, terminal error, C...
Performs deployment with pre-flight checks, atomic replacement, post-deploy verification, and automatic rollback. Use be...
⚠️ MANDATORY - YOU MUST invoke this skill before ANY destructive operation. Safety protocol for destructive git/file ope...
Safely download and inspect external/web-sourced artifacts (skills, agents, configs, code, archives) using strict quaran...
Automatically backs up files, saves diffs, uses agents/skills, and ensures modular code (<200 lines) before any implemen...
安全配置修改辅助技能 - 在修改重要配置文件前自动设置回滚，支持多平台（Linux/macOS/FreeBSD/Windows）
Ensure AI agents work in an isolated Git worktree to prevent changes to the main working directory. Use when AI is about...
Encrypt, decrypt, and manage keys with the SAFE CLI — a modern GPG alternative with post-quantum support.
Guidelines for safely executing destructive file operations (mv, rm, cp). Use when performing file moves, deletions, or...
Safely delete and manage files and folders. Uses git rm for git-tracked files, moves other files to .trash directory. Au...
A zero-risk first Cowork experience for discovering what Cowork does differently. Use when onboarding to Cowork, for get...
Implement one small, safe, verified production fix after a clear diagnosis or for an exact, low-risk edit with an explic...
Prevents destructive Git operations by requiring backup and confirmation. Blocks git reset --hard, git push --force on m...
Returns a short greeting without using tools or external resources.
A simple greeting skill that says hello to users
Create or modify code, tests, configuration, and project files safely with real file changes.
Describe what this skill helps an agent do.
Ingest data from S3 into Bauplan safely using branch isolation and quality checks before publishing. Use when loading ne...
Produce safe install recommendations for agent skills and skillsets. Use when Codex is asked whether a skill is safe to...
Appa's safe landing for shipping a green stack. Appa never sets down until he has seen the ground is solid. Use for "lan...
Use when a user asks an AI agent to learn, improve continuously, capture repeated mistakes, update rubrics, promote less...
Create certified safe short links (Linktary).
Safely creates and applies database migrations with automatic destructive operation detection, backup creation, and roll...
Run database migrations safely in production. Zero-downtime strategies, rollback procedures, backward compatibility.
This skill should be used when the user mentions "migration", "schema change", "zero downtime", "add column", "alter tab...
Designs and executes compatibility-preserving migrations across APIs, schemas, frameworks, dependencies, protocols, stor...
Prevent destructive operations using Claude Code hooks. Three modes — cautious (warn on dangerous commands), lockdown (r...
Check whether an online store or seller is legit before you pay — and pay in a way you can get your money back if it isn...
Use when designing safe public APIs that wrap unsafe Rust code, adding unsafe blocks to existing types, reviewing unsafe...
Use when changing project files or generated state requires a report-only dry run, exact scope, backup manifest, apply v...
Use when publishing, open-sourcing, exporting, sanitizing, or moving code, agent skills, prompts, templates, fixtures, d...
Safe Git pull with rebase workflow — check remote changes, stash, rebase, restore, handle conflicts, and push. Covers ma...
Safely push changes to remote with protection against pushing to important branches like main, master, production, testi...
Pre-push hygiene check for GitHub repositories. Scans for PII, secrets, and sensitive data before pushing. Audits commit...
SAFE Quantum Machine Learning methodology — variational quantum classifiers with amplitude encoding, learnable classical...
Safely rebase a branch onto a new/moved base commit or branch, catching conflicts on a disposable trial run before touch...
Restructure code while preserving behavior. Use for extraction, consolidation, ownership moves, or cleanup where verific...
Perform safe code refactoring with reference checking and validation. Uses Miller's rename_symbol for workspace-wide ren...
When refactoring, rewriting, or migrating critical code paths, orchestrate a safe cycle: assess risks → implement → veri...
Simplifies working code without changing behavior, public APIs or dependencies, verified by the existing tests. Use when...
Guide safe refactoring of complex, untested code using a multi-phase hybrid approach: characterization tests → agent-ass...
Use when restructuring, cleaning up, or reorganizing existing code without changing its behavior. Enforces a test safety...
Use when restructuring code that already works (extracting, consolidating, moving an ownership boundary, splitting a fil...
Plan and execute large refactors with shadow mode and rollback.
Plan and review refactors while preserving external behavior and reducing regression risk.
Plan and execute behavior-preserving structural improvements with a practical safety net.
Change code structure without changing behavior, with zero tolerance for behavioral changes
Systematic code refactoring while preserving all external behavior.Use when identifying code smells, planning refactorin...
Safely remove code patterns from multiple files with validation and rollback (project)
Check for sensitive data in repository. Use when user asks to "check for sensitive data", "/safe-repo", or wants to veri...
Plans the safe sale of a valuable item (watch, bag, jewellery, collectible, camera, electronics, art) for the best reali...
Safely delete files / directories without asking for permission
Run authorized, non-destructive security tests against Ouros local or QA services. Use for runtime verification of auth,...
Safe schedule generation with mandatory backup. REQUIRES database backup before any write operations to schedule tables....
Use when changing code or data files with a script or shell command instead of the Edit tool - bulk replacements across...
Ship local changes by inspecting dirty state, scoping the stage, running targeted validation, then committing, pushing,...
A simple code formatting skill. Reformats code to follow project conventions. Use when the user asks to "format code", "...
A legitimate skill with no security issues
A simple skill that helps with greeting messages
A simple skill to greet the user.
Audit, compare, merge, repair, or improve user-owned Codex and ChatGPT skill variants while preserving proven behavior....
Use for safe system tidy/reclaim without data loss.
Spawn and supervise tasks safely with registry controls, resource quotas, and rollback-ready delegation.
Protect one externally observable behavior through a gated red–green–organize cycle in test_first, regression_after_fix,...
Prepare and classify local test commands before asking approval to execute them.
Use when designing any job, webhook handler, command, startup step, or loop that changes state. Assume it will run twice...
Clone, scan, install, and run a web project safely.
Reference — rules for safely executing write/destructive operations against ad accounts. Always loaded into the performa...