lawzava
UserSkills, plugins, and hooks that change how coding agents work — Claude Code, Codex, OpenCode, Antigravity. Measured, not vibed.
Categories
Indexed Skills (41)
memory-hygiene
Use when auditing, pruning, correcting, refreshing, or automatically applying approved evidence-based cleanup to native harness memory.
designing-frontends
Use to build or redesign a rendered UI when visual direction, palette, typography, layout, motion, or UX copy matters, especially if output looks templated or AI generated.
golang-patterns
Use for Go in an existing project when choosing interfaces, dependency injection, goroutines, channels, context, errors, functional options, or package layout. Skip mechanical edits.
greenfield-go-stack
Use to start or scaffold a Go project, service, SaaS, backend, or server rendered app, or choose its framework, database, auth, payments, hosting, and architecture.
best-of-n
Use when a hard implementation or design task needs independent candidates and one winner chosen by tests or blind comparison. Unlike a council, it selects a work product.
council-adjudication
Use when a hard architecture, design, strategy, or tradeoff decision lacks an executable oracle and needs independent model judgments. Unlike best-of-n, it selects a recommendation.
cross-model-verification
Use when risky billing, auth, concurrency, security, or other high-stakes work needs another vendor to independently refute or adversarially verify it.
effect-broker
Use before a deploy, send, charge, database migration, destructive query, DNS change, external API write, or other costly real-world side effect.
multi-agent-delegation
Use when a scoped build, plan or code review, or visual or browser task should go to a different model or runtime rather than same-model subagents.
orchestrating
Use when two or more independent lanes can run in parallel, an output-only lane protects lead context from raw payload, or a handoff or durable run needs coordination.
wayfinding
Use when long-horizon work has unknown ownership, unresolved decisions, or unclear sequencing that prevents an honest specification or implementation plan.
greenfield-python-stack
Use to start or scaffold a Python project or service, or choose its package manager, tooling, web framework, validation, async, database, middleware, and layout.
python-patterns
Use for Python in an existing project when choosing idiomatic typing, Pydantic models, error handling, asyncio, concurrency, or async design.
greenfield-ts-stack
Use to start or scaffold a TypeScript or Node service, or choose its package manager, tsconfig, tests, linting, validation, web framework, database, and layout.
typescript-patterns
Use for TypeScript in an existing project when choosing idiomatic types, discriminated unions, typed errors, promises, concurrency, or async design.
brainstorming
Use to design a feature, capability, or unclear behavior change. Triggers on "add a feature" or "figure out the approach". Use wayfinding when uncertainty blocks scoping. Skip mechanical edits and bug fixes.
executing-plans
Use to execute a written implementation plan inline with one writer, especially coupled tasks. Use subagent-driven-development for independent tasks with per-task review.
finishing-a-development-branch
Use when verified work needs branch integration or cleanup. Triggers on "work is done", "ship this", "merge or PR", or "wrap up the branch". Code review usually comes first.
project-memory
Use to save or recall durable project knowledge across sessions: decisions, hidden constraints, gotchas, or references. Triggers: "remember this", "note this decision", or "what did we decide". Not for run progress.
receiving-code-review
Use to evaluate review feedback before changing code, especially unclear or suspect suggestions. Triggers on reviewer comments or requested changes. Not for requesting review.
requesting-code-review
Use when finished work needs review before merge. Triggers on "review this", "ready to merge", or "check my work". Verify behavior first.
subagent-driven-development
Use when a plan has independent tasks for subagents or requests recursive multi-writer execution. Triggers on "subagent per task", "fan out plan tasks", or "multi-writer". Use executing-plans for inline work.
test-driven-development
Use for any feature or bug fix. Triggers on "TDD", "test-first", "write the test first", "red-green-refactor", "implement and test", or requests to add code with tests. Diagnose unknown failures first.
using-git-worktrees
Use when starting feature work or an implementation plan that should verify or establish an isolated workspace. Reuse an existing worktree when suitable.
verification-before-completion
Use before claiming work is complete, fixed, passing, ready to merge or publish, and before commits or pull requests. Triggers on any success or status claim.
writing-plans
Use when requirements need a multi-step plan before coding. Triggers on "write a plan", "break into steps", "save the plan", or "do not implement yet". Brainstorm unclear work first.
writing-skills
Use when creating new skills, editing existing skills, or verifying skills work before deployment
upgrading-megapowers
Use when asked to update, upgrade, refresh, reinstall, or migrate Megapowers, check for a newer release, or repair stale installed plugin state.
design-and-plan
Use to specify non-trivial behavior or interfaces, resolve requirements or tradeoffs, or produce a multi-step implementation plan. Do not use for a mechanical edit, routine status update, or execution of a settled plan.
independent-review
Use when security, auth, billing, concurrency, data integrity, or another high-stakes artifact needs adversarial review by a provider other than its author.
test-first-implementation
Use when adding or changing behavior, fixing a confirmed bug, refactoring executable code, or implementing a feature with tests.
verify-and-finish
Use to verify a completed outcome before claiming success, handing off, committing, merging, publishing, deploying, or closing a task. Do not use for an intermediate update, an isolated test run, or merely naming an oracle command.
writing-agent-instructions
Use when creating or revising an agent skill, AGENTS.md, CLAUDE.md, or nested project instructions, including trigger design, progressive disclosure, harness-aware scope, and behavioral validation.
autonomous-run
Use when an approved goal must continue unattended across many steps, milestones, context resets, or sessions.
using-megapowers
Use when starting any conversation - establishes how to find and use skills, requiring skill invocation before any response including clarifying questions
humanizing-prose
Use when the task itself is to draft, rewrite, edit, or preserve human-facing prose for clarity, attribution, or proportion; do not select merely because other work ends with a response.
systematic-debugging
Use when a bug, failing or flaky test, production incident, unexpected output, or performance regression has an unknown cause.
evidence-research
Use to research a decision that needs evidence beyond the repository, such as current upstream behavior, historical rationale, cross-system incidents, or contested claims. Do not use for repository-only code tracing or a summary of provided material.
grill-me
Use when the user asks to be grilled, or wants a plan, decision, idea, or draft stress-tested through a structured interview before any action.
mcp-setup
Use when installing, configuring, debugging, or verifying an MCP server for a harness, or when a configured server's tools are missing or failing.
safe-effects
Use when preparing a deploy, message, charge, migration, destructive query, DNS change, external API write, or any action with a real-world side effect.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.