affaan-m

video-editing

Use it when:: You have raw recordings to cut into a vlog, tutorial, demo or short clips, rather than a video to generate from a prompt.
It does:: Has the agent plan cuts from the transcript, then script FFmpeg trims, Remotion overlays and only the AI assets the footage lacks.

workflow local cli video audio ffmpeg remotion elevenlabs

Age: 5 months
Stars: 300K; #5
Total Downloads: 8.5K; bottom 42%
Weekly Downloads: 1.1K; top 28%

security-review

Use it when:: You're adding authentication, user input, secrets, API endpoints or payment features to a web app.
It does:: Walks ten areas, from secrets and input validation to CSRF, rate limiting and dependencies, each with wrong and right code and checkboxes.

guidelines none none security auth typescript nextjs supabase

Age: 5 months
Stars: 300K; #5
Total Downloads: 7.5K; bottom 38%
Weekly Downloads: 606; top 32%

backend-patterns

Use it when:: You're building or reviewing API routes and their data access in a Node.js, Express or Next.js backend.
It does:: Gives the agent patterns for REST URLs, repository and service layers, middleware, queries, transactions, caching, errors, auth, rate limits, job queues and logging.

guidelines none none backend architecture databases typescript node nextjs supabase

Age: 5 months
Stars: 300K; #5
Total Downloads: 7.1K; bottom 35%
Weekly Downloads: 536; top 33%

coding-standards

Use it when:: You're writing, reviewing or refactoring code and want a shared floor of naming, readability and quality rules.
It does:: States KISS, DRY and YAGNI, then shows good and bad TypeScript for naming, immutability, errors, async code, types, React components and APIs.

guidelines none none code-quality code-review refactoring typescript react

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.9K; bottom 33%
Weekly Downloads: 579; top 32%

market-research

Use it when:: You need market sizing, a competitor comparison, an investor dossier or a technology scan to make a business decision.
It does:: Sources every important claim, flags stale data, includes contrarian evidence and keeps fact, inference and recommendation apart.

workflow none none research marketing finance

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.8K; bottom 33%
Weekly Downloads: 647; top 31%

api-design

Use it when:: You're designing new REST endpoints or reviewing an existing API contract.
It does:: Sets rules for URLs, methods, status codes, response and error shapes, pagination, filtering, sorting, auth, rate limits and versioning.

guidelines none none backend architecture

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.6K; bottom 32%
Weekly Downloads: 519; top 34%

frontend-slides

Use it when:: You want a talk, pitch or workshop deck as a web page, or to turn a .ppt or .pptx file into one.
It does:: Asks the deck's purpose, length and the feeling it should create, generates three one-slide style previews, then builds the deck in the style you pick.

workflow local none slides design animation

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.4K; bottom 31%
Weekly Downloads: 464; top 37%

codebase-onboarding

Use it when:: You join a new project, open a repo in Claude Code for the first time, or want a CLAUDE.md written for it.
It does:: Scans manifests, framework configs, entry points, the top two directory levels, tooling, tests and git history, without reading every file.

workflow local none learning architecture docs-writing claude

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.3K; bottom 30%
Weekly Downloads: 517; top 34%

frontend-design-direction

Use it when:: You're building or polishing a web page, dashboard, app or component that works but looks generic or wrong for its audience.
It does:: Picks purpose, audience, tone and one memorable detail before coding, then records the evidence behind each choice and every content state.

workflow local browser design anti-slop accessibility

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.3K; bottom 30%
Weekly Downloads: 488; top 35%

tdd-workflow

Use it when:: You're adding a feature, fixing a bug or refactoring, optionally starting from a *.plan.md implementation plan.
It does:: Writes user journeys and failing tests, proves the red run, makes the smallest fix, proves green, refactors, then checks 80% coverage.

workflow local none tdd testing git verification typescript node

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.3K; bottom 30%
Weekly Downloads: 543; top 33%

prompt-optimizer

Use it when:: You have a draft prompt for Claude Code and want it sharper, or don't know how to ask ECC for a task.
It does:: Detects your stack, classifies the task's intent and scope, finds missing context and matches ECC commands, skills and agents.

workflow none none requirements agent-workflows skill-discovery

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.3K; bottom 30%
Weekly Downloads: 492; top 35%

database-migrations

Use it when:: You're writing a schema or data migration, adding a column or index to a large table, or planning a zero-downtime deploy.
It does:: Applies five rules, such as forward-only production changes and schema kept apart from data, plus a seven-item safety checklist.

guidelines none none migration databases deployment postgres prisma drizzle kysely django golang-migrate

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.3K; bottom 30%
Weekly Downloads: 544; top 33%

article-writing

Use it when:: You want a blog post, guide, tutorial, essay, launch post or newsletter longer than a paragraph.
It does:: Outlines with one job per section, opens each section with proof or an example, and deletes banned AI phrases.

workflow none none writing anti-slop

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.2K; bottom 28%
Weekly Downloads: 475; top 36%

verification-loop

Use it when:: You've finished a feature or refactor and want proof it's ready before opening a PR.
It does:: Runs the build, tsc or pyright, the linter or ruff, tests with coverage, a grep for keys and console.log, and a diff review.

workflow local none verification testing code-review typescript python

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.2K; bottom 28%
Weekly Downloads: 549; top 33%

deep-research

Use it when:: You want a deep dive, competitive analysis, technology evaluation, market sizing or due diligence rather than a quick answer.
It does:: Asks your goal, splits the topic into 3-5 sub-questions, searches each several ways, then reads 3-5 key sources in full.

workflow local mcp account research web-search firecrawl exa

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.1K; bottom 28%
Weekly Downloads: 494; top 35%

blender-motion-state-inspection

Use it when:: A Blender character looks twisted, mirrored, offset or foot-sliding, or you need to know whether a retargeted motion matches the expected pose.
It does:: Inventories the scene, maps the skeleton, works out facing axes, samples key frames and measures contact, sliding, crossover, twist and scale drift.

workflow none cli animation debugging verification blender

Age: 5 months
Stars: 300K; #5
Total Downloads: 6.1K; bottom 28%
Weekly Downloads: 513; top 34%

context-budget

Use it when:: You ask how much context is left, or want to know how much context your installed skills, agents and MCP tools take up.
It does:: By default, only names the native meter, /context in Claude Code or /status in Codex, without reading files or running tools.

command none none token-saving agent-workflows claude codex

Age: 5 months
Stars: 300K; #5
Total Downloads: 6K; bottom 26%
Weekly Downloads: 504; top 34%

seo

Use it when:: You want better search visibility for a site: an SEO audit, schema markup, sitemap or robots work, or keyword mapping.
It does:: Checks technical blockers before content, then on-page rules, keyword-to-URL mapping and internal links, each with concrete targets.

guidelines none none seo performance

Age: 5 months
Stars: 300K; #5
Total Downloads: 6K; bottom 26%
Weekly Downloads: 467; top 37%

accessibility

Use it when:: You're specifying, building or auditing UI components for keyboard, screen reader and WCAG 2.2 AA support.
It does:: Takes each component from its native role through perceivable, operable, understandable and robust checks, with concrete thresholds.

guidelines none none accessibility mobile

Age: 5 months
Stars: 300K; #5
Total Downloads: 6K; bottom 26%
Weekly Downloads: 483; top 35%

strategic-compact

Use it when:: You run long Claude Code sessions near the context limit and would rather compact between phases than mid-task.
It does:: Gives a table of when to compact, yes after research, planning or debugging, never mid-implementation, and lists what survives compaction.

guidelines none claude-code token-saving agent-workflows claude

Age: 5 months
Stars: 300K; #5
Total Downloads: 6K; bottom 26%
Weekly Downloads: 504; top 34%

design-system

Use it when:: You're starting a design system, checking visual consistency before a redesign, or reviewing a PR that touches styling.
It does:: In one of three modes: pulls tokens from your CSS, Tailwind or styled-components, scores the UI, or checks it for AI-slop patterns.

workflow local mcp browser design-system anti-slop design

Age: 5 months
Stars: 300K; #5
Total Downloads: 6K; bottom 26%
Weekly Downloads: 474; top 36%

browser-qa

Use it when:: A feature just reached staging or preview, or a frontend PR needs checking before it ships.
It does:: Drives the pages through Claude in Chrome, Playwright or Puppeteer in four phases: smoke test, interactions, visual regression and accessibility.

workflow remote mcp browser testing browser-automation verification accessibility performance playwright puppeteer

Age: 5 months
Stars: 300K; #5
Total Downloads: 6K; bottom 26%
Weekly Downloads: 503; top 34%

git-workflow

Use it when:: You're picking a branching strategy, setting commit conventions, deciding between merge and rebase, or resolving conflicts.
It does:: Compares GitHub Flow, trunk-based development and GitFlow by team size and release cadence, then sets conventions for commits, branches, PRs and tags.

guidelines none none git code-review git

Age: 5 months
Stars: 300K; #5
Total Downloads: 6K; bottom 26%
Weekly Downloads: 470; top 37%

search-first

Use it when:: You're about to add a feature, dependency, integration or utility that may already exist.
It does:: Checks which search channels it has, defines the need, searches registries, MCP servers, skills and GitHub, and scores the candidates.

workflow local subagents research implementation architecture

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.9K; bottom 24%
Weekly Downloads: 504; top 34%

brand-voice

Use it when:: You want content or outreach for X, LinkedIn, email or launches to sound like a specific person, not generic AI copy.
It does:: Gathers 5-20 real samples, recent X posts first, and pulls out rhythm, density, capitalization, parentheticals, question use and claim style.

workflow none none writing copywriting anti-slop social-media x-twitter linkedin

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.9K; bottom 24%
Weekly Downloads: 473; top 36%

architecture-decision-records

Use it when:: You've just chosen between frameworks, databases or patterns, or someone asks why the code is built the way it is.
It does:: Spots decision moments in the conversation, drafts a record with context, rejected alternatives and consequences, and writes it once you approve.

workflow local none architecture docs-writing

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.8K; bottom 23%
Weekly Downloads: 476; top 36%

deployment-patterns

Use it when:: You're setting up CI/CD, containerizing an app, choosing a rollout strategy or checking a release is ready for production.
It does:: Weighs rolling, blue-green and canary deploys, and gives Dockerfile, GitHub Actions, health check and Kubernetes probe templates.

guidelines none none deployment ci infrastructure docker github kubernetes

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.8K; bottom 23%
Weekly Downloads: 472; top 36%

hexagonal-architecture

Use it when:: Domain logic is tangled with I/O, or one use case needs several entry points such as HTTP, CLI, queue workers and cron.
It does:: Has the agent put each use case behind ports, interfaces for what it offers and needs, with adapters at the edges and one place that wires them.

guidelines none none architecture refactoring backend typescript

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.8K; bottom 23%
Weekly Downloads: 470; top 37%

frontend-a11y

Use it when:: You build or review forms, modals, dropdowns, tabs or tooltips in React, or a linter or code review flags accessibility problems.
It does:: Shows wrong and right TSX for the issues reviews flag most: unlinked labels, clickable divs, misused ARIA and missing keyboard support.

guidelines none none accessibility react

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.8K; bottom 23%
Weekly Downloads: 471; top 37%

security-scan

Use it when:: You set up or change a .claude/ directory, onboard to a repo with Claude Code configs, or want a periodic security check.
It does:: Runs AgentShield over CLAUDE.md, settings.json, mcp.json, hooks and agent files, and grades the setup from A to F.

tool local cli official security guardrails agentshield claude

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.7K; bottom 21%
Weekly Downloads: 471; top 37%

agentic-engineering

Use it when:: Agents will do most of a piece of engineering work end to end, while humans keep control of quality and risk.
It does:: Has the agent set done criteria and evals first, split work into 15-minute verifiable units, and pick Haiku, Sonnet or Opus by task.

guidelines none none agent-workflows verification planning claude

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.7K; bottom 21%
Weekly Downloads: 456; top 38%

ui-demo

Use it when:: You need a walkthrough, tutorial or feature demo video of a web app for docs, onboarding or stakeholders.
It does:: Maps each page's real fields, rehearses until every selector resolves, then records with a gliding cursor, visible typing, smooth scrolling and step subtitles.

workflow local cli browser video browser-automation playwright

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.7K; bottom 21%
Weekly Downloads: 482; top 35%

agent-architecture-audit

Use it when:: An agent or LLM feature works in the playground but misbehaves in your app and you don't know which layer is at fault, or before shipping one.
It does:: Scopes the system, searches its code, logs, prompts and memory files for known failure patterns, and ties each finding to one of 12 layers.

workflow none none agent-development debugging

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.7K; bottom 21%
Weekly Downloads: 457; top 38%

error-handling

Use it when:: You're designing error types, adding retries, wiring API error responses or writing the messages users see when something fails.
It does:: Gives typed error classes and a Result type in TypeScript, an exception hierarchy with FastAPI handlers in Python, and wrapped sentinel errors in Go.

guidelines none none code-quality backend typescript python go

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.7K; bottom 21%
Weekly Downloads: 474; top 36%

eval-harness

Use it when:: You want pass/fail criteria for an agent task before it starts, or proof that a prompt, agent or model change didn't break what worked.
It does:: Has the agent write capability and regression evals first, implement, run them with code, rule, model or human graders, then report.

workflow local none verification testing agent-workflows claude

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.7K; bottom 21%
Weekly Downloads: 462; top 38%

content-engine

Use it when:: You want posts, threads, video scripts or a newsletter built from something you already have, such as an article, demo, changelog or transcript.
It does:: Pulls 3 to 7 claims from the source, ranks them, gives each output one idea, and adapts the format per platform but not the voice.

workflow none none social-media copywriting writing x-twitter linkedin tiktok youtube

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.7K; bottom 21%
Weekly Downloads: 457; top 38%

dashboard-builder

Use it when:: You're turning a list of metrics into a Grafana, SigNoz or similar dashboard that people will run a service from.
It does:: Starts from operator questions, studies your existing dashboards' JSON and query language, then builds the smallest board that answers them.

workflow local none observability grafana signoz

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.7K; bottom 21%
Weekly Downloads: 449; top 39%

investor-materials

Use it when:: You're preparing a pitch deck, investor memo, one-pager, financial model, use-of-funds table or accelerator application.
It does:: Fixes one source of truth for traction, pricing, raise, use of funds, team and milestones, and stops when numbers conflict.

workflow none none finance slides writing

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.6K; bottom 19%
Weekly Downloads: 458; top 38%

ai-regression-testing

Use it when:: An AI agent both wrote and reviewed a change, or a bug it fixed keeps coming back.
It does:: Runs the test suite and build before any AI review, then writes a regression test for each bug fixed, named after the bug.

workflow local none testing code-review agent-workflows vitest nextjs typescript supabase

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.6K; bottom 19%
Weekly Downloads: 462; top 38%

manim-video

Use it when:: You want a short animated explainer of a graph, workflow, architecture or metric, rather than a talking-head video.
It does:: Writes a one-sentence visual thesis, splits it into 3 to 6 scenes that each prove one thing, then renders a low-quality test before polishing.

workflow local cli video animation diagrams manim

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.6K; bottom 19%
Weekly Downloads: 499; top 34%

agent-introspection-debugging

Use it when:: An agent run loops, hits its tool-call limit, retries without progress, bloats its context or drifts from its task.
It does:: Records the failure, matches it to a known pattern, takes the smallest reversible fix, and only then retries.

workflow none none debugging agent-workflows

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.6K; bottom 19%
Weekly Downloads: 465; top 37%

production-audit

Use it when:: You ask "is this ready to ship?" before a launch, after a merge, or when CI is green but you want the production risks.
It does:: Reads git state, CI, migrations, routes, webhooks and environment docs, then checks security, data, payments, operations and user experience.

workflow none none verification code-review deployment security

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.6K; bottom 19%
Weekly Downloads: 461; top 38%

marketing-campaign

Use it when:: You're planning a product or feature launch across several channels, or need a full content suite from one brief.
It does:: Researches audience and competitors, locks positioning and a campaign angle, then writes each layer of copy in order.

workflow none none marketing copywriting conversion social-media email

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.6K; bottom 19%
Weekly Downloads: 447; top 40%

cost-aware-llm-pipeline

Use it when:: You're building an app or batch job that calls Claude, OpenAI or similar APIs and its spend needs a ceiling.
It does:: Routes small tasks to a cheap model and large ones to a stronger one, tracks spend against a budget, retries only transient errors, and caches long prompts.

guidelines none none agent-development token-saving claude python

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.6K; bottom 19%
Weekly Downloads: 472; top 36%

council

Use it when:: A decision has several credible options and no obvious winner, such as ship now or hold, or a feature flag versus a full rollout.
It does:: Writes down the agent's own position first, then asks Skeptic, Pragmatist and Critic subagents in parallel, each given only the question and the context it needs.

workflow none subagents multi-agent agent-workflows

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.6K; bottom 19%
Weekly Downloads: 458; top 38%

blueprint

Use it when:: A feature, refactor or migration will span several PRs or sessions; it says not to use it for work that fits in one PR.
It does:: Reads the repo, splits the objective into one-PR steps with dependencies, parallel lanes, a model tier and a rollback each, then has a strongest-model subagent review the plan.

workflow local subagents planning handoff multi-agent

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.5K; bottom 16%
Weekly Downloads: 449; top 39%

data-scraper-agent

Use it when:: You want a bot that keeps checking a public site, API or feed, such as job boards, prices, news or GitHub repos, and files what it finds.
It does:: Asks what to collect and where to keep it, then writes a Python project that scrapes, scores items in batches with Gemini and pushes new ones to Notion.

workflow local account scraping agent-development python gemini notion github

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.5K; bottom 16%
Weekly Downloads: 445; top 41%

github-ops

Use it when:: You maintain a GitHub repo and want issues triaged, stale PRs chased, a broken CI run diagnosed or a release cut.
It does:: Classifies issues by type and priority, labels and answers them, checks PRs' CI and mergeability, reads failed logs and creates releases, all with gh.

workflow remote cli issue-tracker issue-tracking ci security github

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.5K; bottom 16%
Weekly Downloads: 450; top 39%

llm-trading-agent-security

Use it when:: You're building or auditing an AI agent that signs transactions, places orders, swaps tokens or touches a treasury.
It does:: Layers independent controls: injection checks on outside data, hard spend limits, simulation before every send, circuit breakers and an isolated hot wallet.

guidelines none none security guardrails web3 finance python ethereum

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.5K; bottom 16%
Weekly Downloads: 475; top 36%

mle-workflow

Use it when:: You're turning notebook code into a production model, or planning, reviewing or debugging a ranking, classifier, forecasting or embedding system.
It does:: Pins the prediction and data contracts first, then builds reproducible training, declares promotion gates before training ends, and plans serving, monitoring and rollback.

workflow local none ml-research deployment observability verification python

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.5K; bottom 16%
Weekly Downloads: 462; top 38%

ai-first-engineering

Use it when:: You're setting process, review gates or ownership rules for a team whose code is largely written by agents.
It does:: Shifts the weight from typing to planning, from anecdotes to eval coverage, and from reviewing syntax to reviewing system behavior.

guidelines none none agent-workflows code-review testing architecture

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.5K; bottom 16%
Weekly Downloads: 446; top 40%

healthcare-phi-compliance

Use it when:: Your code touches patient or clinician data, or you're adding HIPAA- or GDPR-style access controls.
It does:: Sorts data into PHI and other sensitive PII, limits access by facility with row-level security, and logs every access to an insert-only audit table.

guidelines none none security privacy databases postgres supabase

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.5K; bottom 16%
Weekly Downloads: 487; top 35%

security-bounty-hunter

Use it when:: You're preparing a Huntr or HackerOne submission and want findings that pay, not theoretical risks.
It does:: Checks the program's scope, finds real entry points, uses Semgrep only for triage, and follows user input end to end to a dangerous sink.

workflow local none security code-review

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.5K; bottom 16%
Weekly Downloads: 455; top 38%

code-tour

Use it when:: You want an onboarding, architecture, PR, root-cause or security tour a reader can step through file by file.
You get:: A .tours/<persona>-<focus>.tour JSON file in Microsoft's CodeTour format, which opens each step at its file and line, with no change to the source code.

workflow local none docs-writing learning code-review codetour

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.5K; bottom 16%
Weekly Downloads: 449; top 39%

product-capability

Use it when:: Product intent is clear but the rules, invariants and edge cases engineers need still live in someone's head.
It does:: Restates the capability in one sentence, pulls out business rules, invariants, trust boundaries and rollout needs, and marks what's still undecided.

workflow local none requirements architecture handoff planning

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.5K; bottom 16%
Weekly Downloads: 440; top 43%

agentic-os

Use it when:: You want Claude Code to run as a lasting personal or project system with its own agents, memory and scheduled jobs, not one-off chats.
It does:: Lays out a CLAUDE.md "kernel" that routes requests to specialist agent files, plus slash commands, scripts and a data folder.

guidelines none claude-code multi-agent memory agent-workflows architecture claude

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.5K; bottom 16%
Weekly Downloads: 447; top 40%

iterative-retrieval

Use it when:: A subagent needs codebase context you can't predict up front, or agent tasks fail from too much or too little context.
It does:: Searches broadly, scores each file's relevance from 0 to 1, notes what's still missing, refines the search and repeats, at most three times.

workflow none none multi-agent token-saving agent-workflows

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 447; top 40%

regex-vs-llm-structured-text

Use it when:: You're extracting records from text that follows a pattern, such as exam questions, forms, invoices or tables, and LLM cost matters.
It does:: Parses with regex first, scores each item's confidence, and sends only the items below 0.95 to the cheapest LLM for correction.

guidelines none none architecture implementation python

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 447; top 40%

investor-outreach

Use it when:: You're writing to angels, VCs, strategic investors or accelerators: a cold email, a warm intro request, a follow-up or an update.
It does:: Builds each message around why this investor, what the company does with its proof, and one concrete ask, and cuts stock phrases like "I'd love to connect".

guidelines none none email copywriting writing finance

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 440; top 43%

agent-eval

Use it when:: You're choosing between coding agents, or want numbers on whether a model or setup change actually helped.
It does:: Runs each agent several times on tasks you define in YAML, each run in its own git worktree, and judges the result with tests, grep or an LLM.

tool local cli testing agent-development verification agent-eval

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 449; top 39%

crosspost

Use it when:: You want one launch, update or essay out on several social platforms without pasting the same text everywhere.
You get:: A primary version, a variant per platform, a note on what changed and why, and any publishing constraint you still need to settle.

workflow none none social-media copywriting writing x-twitter linkedin threads bluesky

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 442; top 42%

skill-comply

Use it when:: You've written a skill or rule and want to know whether the agent follows it, rather than assume it does.
It does:: Derives the expected steps from the file, writes scenarios at three prompt strictness levels, runs each with claude -p and classifies the tool calls.

tool local cli claude-code skill-authoring testing verification claude

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 443; top 41%

click-path-audit

Use it when:: Users report a button that does nothing though debugging found no bug, or after changing a shared state store.
It does:: Maps what every store action sets and resets, then traces each handler's calls in order against what its label promises.

workflow none none debugging code-review react zustand

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 448; top 40%

benchmark

Use it when:: You want a PR's performance impact, baselines before a launch, or numbers behind an "it feels slow" report.
It does:: Measures LCP, CLS, INP and page weight in a browser, p50 to p99 latency over 100 calls per endpoint, and build, test and lint times.

workflow local mcp browser performance testing ci

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 446; top 40%

healthcare-emr-patterns

Use it when:: You're building patient encounters, prescriptions, lab displays or clinical decision support into a medical records system.
It does:: Judges each design by "could this harm a patient?", from interaction checks that block prescribing to alerts that can't be dismissed.

guidelines none none architecture accessibility design

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 482; top 35%

safety-guard

Use it when:: An agent works on production systems, runs unattended, or should only touch one folder during a migration or deploy.
It does:: Careful mode pauses before rm -rf, force pushes, DROP TABLE and similar, and suggests a safer command; Freeze mode limits writes to one directory.

mode none none guardrails scope-control

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 451; top 39%

product-lens

Use it when:: You're about to start a feature, are stuck choosing between ideas, or want a sanity check before a launch.
It does:: Asks who it's for, the pain, why now, the MVP and the anti-goal; or scores your repo's product-market-fit signals; or times onboarding as a new user.

workflow local none requirements planning interview

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 445; top 41%

agent-payment-x402

Use it when:: An agent has to pay for an API, a service or another agent itself, or your API should charge agents per call.
It does:: Picks agentwallet-sdk, OKX's Agent Payments Protocol or the upstream x402 packages by chain and side, then wires spending controls that fail closed.

guidelines none none payments web3 agent-development guardrails x402 solana okx

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 443; top 41%

api-connector-builder

Use it when:: You're adding one more integration, such as a Jira or Slack provider, to a repo that already has several.
It does:: Studies at least two existing connectors first, narrows the target API to what the repo needs, then builds it in the same layers.

workflow local none implementation backend code-quality

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 443; top 41%

knowledge-ops

Use it when:: You say "save this to the KB" or "what do I know about X", or want notes, sessions and plans in one canonical place.
It does:: Classifies each piece of knowledge, searches every layer for an existing entry, then stores it in the right layer and updates indexes.

workflow remote mcp claude-code issue-tracker memory notes issue-tracking github linear

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 441; top 43%

terminal-ops

Use it when:: You ask the agent to run something, check a repo, fix a failing build or test, or push a fix.
It does:: Pins down repo, branch, diff and mode, reads the error before editing, fixes one failure at a time and reruns the proving command.

workflow remote none verification debugging ci scope-control git

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 442; top 42%

hipaa-compliance

Use it when:: A task mentions HIPAA, protected health information (PHI), covered entities or business associate agreements (BAAs) in US healthcare software.
It does:: Asks before PHI moves: is it PHI, is this a covered entity or business associate, does the vendor need a BAA, is access minimal and audited.

guidelines none none privacy security

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 478; top 35%

research-ops

Use it when:: You ask for something current, a comparison, a look-up on people or companies, or a question you'll keep asking.
It does:: Sorts the ask into a quick fact, a comparison, an enrichment pass or a monitoring candidate, then takes the lightest search that answers it.

workflow none none research web-search

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 446; top 40%

homelab-network-setup

Use it when:: You're planning a home or lab network, outgrowing the ISP router, or fighting double NAT, flaky Wi-Fi or servers that change address.
It does:: Separates gateway, switch and access point roles, compares gateway options and lays out an IP plan that's ready for VLANs later.

guidelines none none infrastructure personal

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 449; top 39%

ios-icon-gen

Use it when:: An iOS or macOS app needs new icons in its asset catalog, or placeholder icons replaced with ones that match.
It does:: Searches Iconify's 275,000+ open-source icons or picks an SF Symbol, then renders it at your size and colour with bundled scripts.

tool local cli image mobile xcode iconify sf-symbols

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 445; top 41%

skill-stocktake

Use it when:: Your ~/.claude/skills folder has grown and you want to know which skills still earn their place.
It does:: Inventories global and project skills, has subagents judge about 20 at a time for overlap, staleness and usefulness, and saves the verdicts.

workflow local subagents cli claude-code skill-authoring self-improvement claude

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 449; top 39%

skill-scout

Use it when:: You're about to create a skill, or ask whether a skill for some job already exists.
It does:: Searches your installed and marketplace skills, then GitHub and the web, reads each outside match for risky commands, and ranks up to ten.

workflow none none skill-discovery skill-security claude github

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 441; top 43%

enterprise-agent-ops

Use it when:: Agents run continuously or in the cloud, beyond single CLI sessions, and need production controls.
It does:: Names four areas to cover (lifecycle, observability, safety, change management), the baseline controls and five metrics to track.

guidelines none none observability deployment guardrails

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 439; top 44%

team-builder

Use it when:: You have a collection of agent persona files and want several specialists, such as security and SEO, to look at one task.
It does:: Lists your agents by domain, takes your picks by number or name, then runs each as a parallel subagent with its persona and your task.

workflow none subagents claude-code multi-agent agent-workflows

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 445; top 41%

santa-method

Use it when:: Output will be published, deployed or seen by users, or must meet compliance, brand or accuracy rules; not for drafts or what tests already check.
It does:: Spawns two reviewer subagents with the same rubric and no shared context; if either fails the output, all flagged issues get fixed and fresh reviewers check again.

workflow local subagents verification code-review multi-agent

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 445; top 41%

canary-watch

Use it when:: You've just deployed, merged a risky PR or upgraded dependencies and want to confirm the live site didn't regress.
It does:: Checks eight things, from HTTP status and new console errors to LCP, CLS and INP, key page elements, API response times, static assets and SSE heartbeats.

workflow local browser deployment observability performance verification

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 444; top 41%

healthcare-cdss-patterns

Use it when:: You're building drug interaction checks, dose validation, clinical scores such as NEWS2 or qSOFA, or safety alerts inside an electronic medical record.
It does:: Has the agent write the decision engine as pure functions with no side effects, which take clinical data and return alerts.

guidelines none none implementation testing typescript

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 478; top 35%

healthcare-eval-harness

Use it when:: You're about to deploy an EMR or EHR app, or changed its clinical logic, patient-data schema or access control.
It does:: Runs five test suites in order; one failure in the three critical ones blocks the deploy, and the other two warn below 95%.

workflow none none testing ci verification privacy jest github

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 477; top 36%

finance-billing-ops

Use it when:: You want a sales or refund snapshot, a pricing comparison, or proof of whether team billing, seats or quota stacking work in code.
It does:: Starts from live billing data or a dated snapshot, separates customer incidents from product questions, and reads the checkout, pricing and entitlement code.

workflow none account finance payments stripe

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 445; top 41%

lead-intelligence

Use it when:: You want a ranked list of people to reach for sales, partnerships or fundraising, and the warmest way to reach each one.
It does:: Scores people on role, industry, recent activity, reach, location and engagement, then ranks your mutual connections by how many targets they bridge to.

workflow remote mcp account subagents research marketing email social-media exa x-twitter linkedin dual-use

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 441; top 43%

workspace-surface-audit

Use it when:: You're setting up Claude Code or Codex and want to know what your workspace can already do and what to add.
It does:: Inventories harness configs, MCP and LSP servers, plugins and connectors, and services implied by .env key names, never printing secret values.

workflow none none skill-discovery agent-workflows claude codex

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 441; top 43%

automation-audit-ops

Use it when:: Hooks, cron jobs, GitHub Actions, MCP servers and connectors have piled up and you don't know which are live, broken or redundant.
It does:: Reads the real configs and logs, labels each item configured, authenticated, recently verified, stale or missing, and backs each claim with a file, run or log.

workflow none none agent-workflows ci verification github mcp

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 439; top 44%

email-ops

Use it when:: You want the agent to triage your inbox, draft or send a reply, or check whether a message really went out.
It does:: Settles the account, thread and whether you want a draft or a live send, reads the thread first, then drafts, or sends when you ask.

workflow remote account email guardrails

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 440; top 43%

homelab-network-readiness

Use it when:: You're about to split a flat home network into VLANs, point DHCP at Pi-hole or AdGuard Home, or add VPN access.
It does:: Takes an inventory of your edge, gateway, switches, Wi-Fi, subnets, DNS and DHCP, and how you'd recover, then looks for ways you could lock yourself out.

workflow none none infrastructure security planning personal pi-hole wireguard

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 442; top 42%

homelab-vlan-segmentation

Use it when:: You want IoT devices, guests and your own computers on separate networks at home, or inter-VLAN rules aren't working.
It does:: Lays out a five-VLAN design with subnets, then gives click paths for UniFi and pfSense/OPNsense and commands for MikroTik.

guidelines none none infrastructure security personal unifi pfsense opnsense mikrotik

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.4K; bottom 13%
Weekly Downloads: 450; top 39%

content-hash-cache-pattern

Use it when:: A pipeline keeps re-parsing the same PDFs, images or text files, or a CLI needs a --cache/--no-cache option.
It does:: Keys each result by a SHA-256 of the file's content, stores it as its own JSON file, and wraps caching around the untouched processing function.

guidelines none none performance architecture python

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 441; top 43%

visa-doc-translate

Use it when:: A visa application needs a bank, income, employment or retirement certificate, or another official document, translated into English.
It does:: Converts and rotates the image, reads it with OCR, translates it keeping every figure, then writes and runs a Python script to build the PDF.

workflow local cli pdf documents personal python reportlab

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 439; top 44%

ralphinho-rfc-pipeline

Use it when:: A feature is too big for one agent pass and you have an RFC to split among several agents.
It does:: Breaks the RFC into a graph of dependent work units, runs each through research, plan, build, tests and review, then merges them through a queue.

workflow local subagents multi-agent planning agent-workflows git

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 439; top 44%

rules-distill

Use it when:: You've installed many skills and want the principles they share written once into your rule files, such as monthly or after installing new skills.
It does:: Scans your skills and rules with scripts, has subagents cross-read clusters of skills, and keeps only principles found in 2+ skills that your rules lack.

workflow local subagents claude-code cli self-improvement memory

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 440; top 43%

customer-billing-ops

Use it when:: A customer says billing is broken, wants a refund or can't cancel, or you're chasing duplicate subscriptions, failed renewals or churn risk.
It does:: Finds the customer from an email or Stripe ID, puts the case in one of five buckets, then prefers restoring self-serve control over refunding.

workflow remote account payments ecommerce stripe

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 439; top 44%

google-workspace-ops

Use it when:: You need to find a plan, tracker or deck in Google Drive and update, clean up, summarize or import it.
It does:: Searches Drive for the exact file and its duplicates, summarizes its structure, then edits by index in Docs, by tab and range in Sheets, or by slide.

workflow remote account documents spreadsheets slides google-workspace

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 440; top 43%

gan-style-harness

Use it when:: You want a whole app or a polished frontend built from a one-line brief and will spend hours and $50–200 on quality.
It does:: A planner expands the brief into a spec, a generator builds it, and an evaluator clicks through the running app, scores it and sends feedback for the next round.

workflow local claude-code subagents mcp browser multi-agent agent-workflows implementation testing playwright

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 439; top 44%

opensource-pipeline

Use it when:: You want to open-source a private project without leaking secrets, personal data or internal references.
It does:: Chains three agents: a forker copies the project to ~/opensource-staging and strips secrets, a sanitizer scans it, a packager writes CLAUDE.md, setup.sh, README, license and contributing files.

workflow remote cli subagents claude-code security privacy git docs-writing github git

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 442; top 42%

connections-optimizer

Use it when:: You want to clean up who you follow on X, rebalance LinkedIn toward your current priorities, or find who to reconnect with.
It does:: Scores each follow or connection on reciprocity, activity, priority fit and bridge value, in a light-pass, default or aggressive mode.

workflow remote account browser social-media messaging email x-twitter linkedin

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 439; top 44%

project-flow-ops

Use it when:: Your open PR and issue backlog needs triage, or you're unsure what deserves a Linear issue and what stays on GitHub.
It does:: Reads each issue or PR's state, reviews, CI and links, classifies it, and creates or updates Linear only for active, delegated or cross-team work.

workflow remote issue-tracker account issue-tracking code-review github linear

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 439; top 44%

social-graph-ranker

Use it when:: You want to know which of your connections can get you a warm intro to a list of people, companies or an ideal customer profile.
It does:: Weights each target, scores each connection by its paths to them, halving the value per extra hop, then adds second-order reach and a responsiveness bonus.

workflow none account social-media marketing research x-twitter linkedin

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 442; top 42%

defi-amm-security

Use it when:: You're writing or auditing an AMM, liquidity pool, LP vault or swap function in Solidity.
It does:: Pairs each vulnerability with vulnerable and hardened code, using OpenZeppelin's ReentrancyGuard, SafeERC20 and Ownable2Step, a Uniswap V3 TWAP read and FullMath.

guidelines none none web3 security solidity openzeppelin slither

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 442; top 42%

unified-notifications-ops

Use it when:: CI failures, review requests and issue updates land in scattered places and the noise gets ignored.
It does:: Inventories event sources, channels, duplicate paths and silent failures, then routes each event family by a four-level severity model.

workflow none none observability issue-tracking ci agent-workflows github linear

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 439; top 44%

messages-ops

Use it when:: You ask the agent to check your texts or DMs, find a code that just arrived, or read a thread before a follow-up.
It does:: Settles the surface, sender, service and time window first, searches only the recent window for a code, and stops once it's found.

workflow none account messaging personal

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 436; top 45%

evm-token-decimals

Use it when:: You read ERC-20 balances or price tokens across EVM chains, or a balance or USD value is off by orders of magnitude.
It does:: Has the agent query decimals() on-chain, cache it per chain and token address, and normalize amounts with exact math.

guidelines none none web3 finance debugging solidity python typescript

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 438; top 45%

network-interface-health

Use it when:: A host or VLAN has packet loss or intermittent reachability, or a switch port shows CRCs, drops, resets or flaps.
It does:: Captures a counter baseline, waits, captures again and compares the increments, on both ends of the link.

workflow none none debugging infrastructure

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 452; top 39%

network-config-validation

Use it when:: You're about to paste a Cisco IOS-style config into a router or switch, or let automation push a generated one.
It does:: Checks destructive commands, management exposure, duplicate IPs and overlapping subnets, undefined ACLs or route-maps, then hygiene such as NTP and logging.

workflow none none infrastructure security verification code-review cisco-ios python

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 449; top 39%

network-bgp-diagnostics

Use it when:: A BGP neighbor is stuck in Idle or Active, a session is up but prefixes are missing, or unexpected prefixes are advertised.
It does:: Pins down the neighbor, address family, VRF and ASNs, reads state and last reset reason, proves reachability, then checks route policy.

workflow none none debugging infrastructure

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 447; top 40%

benchmark-optimization-loop

Use it when:: You're asked to speed something up, try dozens of optimizations, or pick the best implementation by repeated measurement.
It does:: Pins down the operation, metric, correctness gate, baseline and search budget, then benchmarks variants that each test one hypothesis on the same input.

workflow local none performance verification

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 446; top 40%

latency-critical-systems

Use it when:: p95 latency, throughput or data freshness matters: realtime dashboards, market-data feeds, streaming agents, execution gateways, queues or caches.
It does:: Tracks p50, p95 and p99 latency, freshness age, queue depth and cache hit rate separately, and maps the path from source event to what the user sees.

workflow local none performance observability verification

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 449; top 39%

parallel-execution-optimizer

Use it when:: You want a task done much faster through parallel work: concurrent agents, batched tool calls, isolated worktrees or many verification lanes.
It does:: Turns the task into a dependency graph of lanes, marks each parallel, sequential or gated, and runs lanes together only when their writes can't collide.

workflow local none multi-agent agent-workflows verification

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.3K; bottom 9%
Weekly Downloads: 443; top 41%

data-throughput-accelerator

Use it when:: An ingestion, backfill, export, ETL job, warehouse load or table sync is too slow and must get faster without losing data.
It does:: Separates extraction, transfer, load, transform and freshness, measures the backlog, then benchmarks batch sizes, worker counts, warehouse SQL and file grouping.

workflow remote none performance databases verification

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.2K; bottom 6%
Weekly Downloads: 439; top 44%

prediction-market-oracle-research

Use it when:: You're considering prediction markets as a data source, forecasting input or oracle-like signal for a product, agent, dashboard or decision.
It does:: Records market-implied probabilities with timestamps and links, rates liquidity, spread, age and resolution, and compares them with filings, news, polls or internal data.

workflow none none research finance

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.2K; bottom 6%
Weekly Downloads: 439; top 44%

prediction-market-risk-review

Use it when:: A prediction-market, basket, oracle or trading-agent workflow is about to handle venue logins, API keys, portfolio data or trade planning.
It does:: Checks five gates: advice boundary, venue and regulatory limits, data quality, security and privacy.

workflow none none finance security privacy guardrails

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.2K; bottom 6%
Weekly Downloads: 444; top 41%

recursive-decision-ledger

Use it when:: You ask for repeated rollouts, recursive reasoning, stochastic or high-dimensional search, or ensemble comparisons and want a visible evidence trail.
It does:: Loads the prior ledger, adds fresh information, runs a bounded search, and marks each candidate accept, watch, reject, decay watch or needs replay.

workflow local none agent-workflows memory guardrails

Age: 5 months
Stars: 300K; #5
Total Downloads: 5.2K; bottom 6%
Weekly Downloads: 437; top 45%

ecc Skill collection