Start with the job, then find an open-source answer

Check an academic manuscript before submission for fabricated references, incorrect DOIs, bibliographic errors, missing text-to-list matches, and citations that do not support the associated claim. The skill accepts DOCX, TeX, and Bib inputs and uses CrossRef plus web search for cross-checking.

Before custom code is written, the skill makes an agent state the need, search the current repository, package registries, MCP servers, installed tools, and maintained open source options, then document a decision. The final verdict must choose Adopt, Extend, Compose, or Build using evidence such as fit, maintenance, documentation, licensing, and dependencies.

Leave requested changes directly beside the relevant source code using @feedback annotations in common comment styles. The plugin finds unaddressed notes, implements each request, and adds an @agent-response for review; once approved, a resolve command removes both annotation types and leaves clean code ready to commit.

Ask Claude to deploy a frontend project and this skill handles repository creation, GitHub Actions setup, Pages activation, and deployment monitoring through to the live URL. It also fixes common Vite base-path failures automatically. Git and an authenticated GitHub CLI are required, while some frameworks such as CRA, Next.js static export, or Angular may still need their documented configuration changes.

crossfire runs a four-phase actor-critic workflow in which Claude plans and reviews while Codex implements and audits. Task tiers scale from quick edits to architecture-level changes, with deterministic tests, linters, debate, and fix-and-reaudit loops added as risk increases. Both Claude Code and Codex CLI are required, and the documented setup includes Windows-specific defenses.

ai-cli-kit teaches AI coding tools to identify the operating system, shell, container, SSH context, and Python environment before writing commands. It applies platform-specific rules for quoting, variables, line endings, error handling, and unattended execution, and can provide parallel Bash and PowerShell variants when needed. It is distributed as a Claude Skill, Cursor rules, a Claude Code file, and a universal system prompt.

Progressive Estimation sizes individual tasks or entire backlogs by accounting for agent rounds, integration work, human planning, review, and fixes. It produces PERT-based expected and committed estimates, confidence bands, token usage, and optional model costs, with formats for common issue trackers. Teams can calibrate future estimates against actual results. The project is still in early development, so formulas and defaults may change and should be treated as structured guidance rather than certainty.

This skill helps Claude handle ServiceNow development and architecture requests with platform-specific patterns instead of generic JavaScript advice. It separates domain references from coding standards for areas such as GlideRecord, Flow Designer, security, and performance, then produces ServiceNow-specific code, workflow or architecture guidance with relevant role, plugin, testing, and deployment assumptions.

Paste a Polymarket profile URL to analyze public trades, current positions, and order books for the wallet’s most active markets. The report covers realized performance, drawdown, market concentration, buy-versus-sell pricing, round-trip results, holding times, and liquidity signals to help separate repeatable behavior from lucky outcomes. Results are anonymized by default, and the tool neither trades nor copies positions.

Weaver Mind looks across domains such as biology, architecture, game design, jazz, and physics for patterns that may apply to a problem after the obvious playbook has run out. Its report explains why each analogy transfers, how to act on it, the tradeoffs, effort, and potential impact, then ends with a wildcard and the smallest promising experiment you can run within seven days.

Fetch a Databricks job by name, run ID, or workspace URL, then download and decompress driver and executor logs from Unity Catalog Volumes and merge them chronologically with source labels. The CLI supports focused filters and JSONL output. It requires a configured Databricks CLI and does not currently support S3 log destinations.

Import trading records in PDF, CSV, XLSX, or JSON format and generate a self-contained HTML dashboard covering profit and loss, win rate, drawdown, equity curves, and time-based performance. The analysis includes Kelly sizing, risk of ruin, Monte Carlo scenarios, and behavioral patterns such as loss streaks or revenge trading. Parsers cover several named brokers plus heuristic handling for generic CSV files.

Describe an AWS problem such as recent 500 errors, authentication failures, or crashing ECS tasks. The skill selects relevant logs, builds a CloudWatch query, saves the result, and reports likely causes; it can also show health, alarms, and trends. A setup command discovers log groups, ECS resources, alarms, and the active region. Claude Code and a credentialed AWS CLI are required.

Use Claude Code to query Bags tokens, pools, swap quotes, claimable fees, and wallets linked to social handles. The kit can scaffold fee-sharing apps, trading bots, or token launchers, and provides guided workflows for token launches, fee allocation, and performance analysis. Running the bundled tools requires Node.js 18 or later and a Bags API key.

Guide Claude Code through the Python package lifecycle, from a src-layout scaffold and dependency setup to API naming, structured errors, tests, documentation, deprecations, and PyPI releases. The conventions use tools such as uv, pytest, rich, and MkDocs. The skill can also inspect existing projects for common packaging mistakes and configure pre-commit checks.

screencli-skill lets coding agents such as Claude Code, Cursor, and Windsurf drive a browser from a natural-language task and record the session. The resulting demo can include gradient backgrounds, automatic zoom, click highlights, and cursor trails before being uploaded to a shareable link. It is designed for product walkthroughs and requires Node.js 18 or newer plus FFmpeg.

Tvastar scans an existing frontend stack, installed component libraries, and color tokens before proposing a design system and page plan. After user approval, it can generate code and run checks for visual consistency, framework mistakes, motion, performance, and WCAG accessibility. It targets environments such as Next.js, React, and Tailwind, and requires Claude Code with Python 3.10 or newer.

reddit-intel collects posts for a keyword through Reddit's public JSON interface, combines several sort orders, and deduplicates the results; deep mode also retrieves selected comments. It translates and classifies discussions by intent, sentiment, competitors, use cases, and action items, then exports post-level data and a summary sheet to Excel. Because Reddit exposes only a limited result window, the tool is better for fast directional research than complete archives or continuous monitoring.