Latest Agent Skills
Everything added to the registry recently, newest first.
- jeffersonvalente BundleLint And Validate
Lint, type-check and security-scan commands per ecosystem, with a fix-and-rerun loop. Use after changing code, before calling a task done, or when asked to run static analysis.
- jeffersonvalente SkillBehavioral Modes
Working modes (brainstorm, implement, debug, review, teach, ship) and the output style of each. Use to pick how to respond when the kind of request is unclear.
- jeffersonvalente BundleDesign Flow
Designs and iterates on frontend UI through shape, craft, critique, audit and polish, writing production code. Use when the user wants to design, redesign, critique, polish or harden a page, dashboard or component, set a visual direction, or make it bolder or quieter. Not for backend-only tasks.
- jeffersonvalente BundleParallel Agents
Runs Cardume agents singly, in sequence or in parallel and merges the results. Use when subtasks need different expertise or a review needs several perspectives.
- jeffersonvalente BundleDatabase Design
Picks a database and ORM, designs schemas and indexes, plans safe migrations. Use when designing a schema, adding indexes or planning a migration.
- jeffersonvalente BundleAPI Patterns
Chooses REST, GraphQL or tRPC and sets versioning, pagination, auth and rate limits. Use when designing an API or reviewing endpoint contracts.
- jeffersonvalente BundleTesting Patterns
Test pyramid, mocking, fixtures and test organization. Use when designing tests, choosing what to mock or reviewing test quality.
- jeffersonvalente SkillRed Team Tactics
MITRE ATT&CK-based adversary tactics and reporting. Use when planning or reviewing an authorized red team engagement.
- jeffersonvalente BundleMCP Builder
MCP server tool design, schemas, errors and security. Use when creating or reviewing an MCP server or its tool definitions.
- jeffersonvalente BundleBrainstorming
Asks the clarifying questions that shape a vague idea: purpose, users, scope, trade-offs. Use before building a vague "create X" request or a change with unclear scope.
- jeffersonvalente BundleWebapp Testing
E2E strategy, Playwright scripts, and visual and accessibility checks. Use when testing a web app in a browser or auditing a deployed URL.
- jeffersonvalente BundlePlan Writing
Writes a short plan file of 5-10 verifiable tasks with dependencies. Use when breaking down a new project, feature, complex bug fix or multi-file refactor.
- jeffersonvalente BundleArchitecture
Weighs trade-offs and picks patterns (monolith or microservices, layering, events). Use when choosing an architecture or reviewing a system design.
- jeffersonvalente SkillOrchestrate
Runs a two-phase multi-agent workflow on the given task: project-planner writes a plan for approval, then 3+ specialist agents implement or analyze in parallel, verification scripts run, and results are merged into one Orchestration Report.
- jeffersonvalente BundleApp Builder
Scaffolds a new app from a plain request: picks the stack and coordinates specialists. Use when asked to create a full-stack, mobile, desktop, CLI or extension app from scratch, or add a large feature to one.
- jeffersonvalente SkillRust Pro
Rust 2024 with Tokio and axum: async, types, errors, performance. Use when writing, reviewing or optimizing Rust code.
- jeffersonvalente SkillTdd Workflow
TDD red-green-refactor workflow. Use when implementing a feature or bug fix test-first.
- jeffersonvalente SkillDeploy
Runs pre-flight checks and deploys to a target environment (preview, staging, or production), with rollback support. Takes an optional environment or subcommand such as check or rollback.
- jeffersonvalente SkillClean Code
Coding standards (SRP, DRY, KISS, YAGNI, naming) with a pre-edit dependency check and a completion self-check. Use when writing, editing or finishing code.
- jeffersonvalente SkillBrainstorm
Explores a project or feature idea before implementation: clarifies the goal and constraints, lays out at least three approaches with pros, cons, and effort, and recommends one. Produces no code.
- jeffersonvalente SkillBash Linux
Bash patterns and script templates for Linux and macOS. Use when writing Bash scripts or running non-trivial shell commands.
- jeffersonvalente SkillReview Pr
Reviews a data/analytics PR: SQL, grain, cost and DAG risks plus data evidence from a configured non-production environment, then drafts a review the user approves before anything is posted. Use when a data pipeline PR needs review.
- jeffersonvalente SkillStandards
Scans the repository and writes docs/standards.md with its real conventions, each backed by evidence. Use when starting level-2/3 work without that file, or when conventions changed.
- jeffersonvalente SkillImplement
Builds an approved tasks.md one task at a time, with a fresh subagent, a verification command and a separate reviewer. Use to implement level-2/3 work once its tasks are approved.
- jeffersonvalente BundleRelease
Prepares a release: checks the approved gate or explicit OK, detects existing release tooling, picks the semver bump with reasons, writes the Keep a Changelog entry and docs/release/vX.Y.Z.md (notes, rollback, honest announcement), and hands over tag/publish commands for confirmation. Use when a version is ready to ship; invoked by the user (/cardume-core:release) or the release-manager agent.
- jeffersonvalente SkillPreview
Manages the local development preview server via cardume-auto-preview: start, stop, restart, status, or health check, with port-conflict handling.
- jeffersonvalente SkillEnhance
Adds or changes a feature in an existing application: loads project state, plans affected files, asks approval for major changes, applies them via the relevant agents, and refreshes the preview.
- jeffersonvalente SkillVerify
Runs every acceptance criterion of a spec and records evidence in verify.md, listing gaps. Use after implementing a spec, before calling the feature done or opening its PR.
- jeffersonvalente SkillStatus
Shows the current project status board: project info and tech stack, features done and pending, agent progress, file counts, and preview server health.
- jeffersonvalente SkillReview
Reviews a change with parallel specialist reviewers, verifies each finding against the code and flags outdated specs. Use before merging a non-trivial change or when asked for a code review (PR, branch or diff).
- jeffersonvalente SkillCreate
Scaffolds a brand-new application from scratch: asks clarifying questions, has project-planner write a plan, builds it with the app-builder skill and specialist agents, then starts a preview. Not for features in an existing project.
- jeffersonvalente SkillTasks
Breaks an approved spec into ordered tasks in tasks.md, each with an owner and a verification command, optionally opening GitHub issues. Use after a spec is approved, before implementing level-2/3 work.
- jeffersonvalente SkillSetup
Adds or updates the Cardume conventions block in this project's CLAUDE.md through a deterministic script (backup, diff, only the marked block changes; AGENTS.md is imported, never edited) and suggests stack and code-intelligence plugins.
- jeffersonvalente SkillDebug
Investigates the given symptom or error systematically: gathers evidence, ranks hypotheses, tests them one by one, fixes the root cause, and reports it in a structured Debug format with prevention steps.
- jeffersonvalente SkillSpec
Asks the questions that change the design, then writes a spec in docs/specs/ with testable requirements and non-goals. Use for level-2 work (a new feature, or a change to more than ~5 files or more than one area) before any code; not for level 0-1 edits.
- jeffersonvalente SkillPlan
Creates a project plan file (docs/PLAN-{task-slug}.md) for the given request using the project-planner agent: clarifying questions, task breakdown, agent assignments, and a verification checklist. Writes no code.