github.com/affaan-m/ECC295k
The agent harness performance optimization system.
/plugin marketplace add affaan-m/ECC
Diagnose AI agent failures with a reproducible capture and report, not a retry.
Trim a repo's agent setup into daily and library buckets using evidence.
Design REST resources, status codes, pagination, filtering and versioning.
Write long-form articles and guides in a consistent voice from examples.
Apply server architecture and data access patterns in Node, Express and Next.js.
Score competitors on nine weighted dimensions with explicit 1 to 5 rubrics.
Work through structured interviews to define a brand's identity and voice.
Build a writing style profile from real samples and reuse it across content.
Use Bun as runtime, package manager, bundler and test runner, with migration notes.
Analyze a change's impact before making it, then compare it with the diff.
Apply baseline naming, readability and immutability conventions in code review.
Identify and tier a competitor set before any benchmarking begins.
Assemble scored competitor findings into a decision-ready strategy report.
Build platform-native content systems and calendars for social and newsletters.
Keep stable domain terms in CONTEXT.md and route hard decisions to ADRs.
Adapt and distribute content across X, LinkedIn, Threads and Bluesky.
Research a topic across many web sources and deliver a cited report.
Coordinate several agent sessions in parallel panes with dmux.
Audit documentation roles, links and ADR references without changing files.
Fetch current library and framework docs through Context7 instead of memory.
Write Playwright tests with page objects and fix flaky runs in CI.
Run formal evals on a Claude Code workflow before trusting or changing it.
Search the web, code and companies with Exa's neural search engine.
Generate images, video and audio through the fal.ai MCP server.
Apply React and Next.js state and render performance patterns.
Build animated HTML presentations or convert PowerPoint files to web slides.
Draft pitch decks, memos and financial models that stay consistent.
Draft concise cold emails and follow-ups to angels, VCs and accelerators.
Size markets and compare competitors with sourced, decision-oriented summaries.
Flag mutable package references in MCP config without running the servers.
Build MCP servers in TypeScript with tools, resources, prompts and transports.
Run production ML work from data contracts through training, deployment and rollback.
Set up a module's local agent instructions and interface ownership when it changes.
Track downstream consumers in REGRESSION.md and run affected commands after a change.
Speed up Next.js 16 builds and dev with Turbopack, and compare it with webpack.
Open a plan in a browser canvas to annotate, approve or request changes.
Turn PRD intent into an implementation-ready capability plan with open decisions.
Check auth, user input, secrets and payment code against a security checklist.
