Skills, subagents, slash commands, hooks, settings, plugin bundles and MCP servers that install into Claude Code. 381,855 indexed, measured the same four ways as everything else here.
6 Claude skills for the full job hunt — find real openings (role-finder), decode a JD with its subtext (jd-analyzer), plan what to learn (gap-planner), turn messy / spoken experience into role language (experience-mapper), polish it into résumé bullets (bullet-builder), and assemble a tailored résumé (resume-tailor).…
★not rated 4 4mo agoA
tokens not measured
originalMIT
Use when starting or finishing any coding/editing task in a project that follows a documentation-first AI collaboration workflow — declare scope before editing, log every change afterward with a changelog entry, scaffold a brand-new project with this discipline (AGENTS.md/README/docs skeleton), or audit and reorganize…
End-to-end autonomous development pipeline. Goes from idea to implementation via structured specs, dependency-aware task planning, and parallel multi-agent execution.
★not rated 4 6mo agoA
tokens not measured
originalMIT
Process the garden vault. Files inbox captures into atomic notes, adds wiki-links, dedupes, updates MOCs, and commits. Run on a schedule (cron or routine), not by hand. Reads /garden/meta/gardener-rules.md for heuristics.
plumb-line provenance: epistemic honesty enforced by tooling. Five skills give any repository a clear source-truth layer, visible uncertainty, quarantined fakery, reproducible outputs, and rules enforced by machines rather than goodwill. New here? Run plumb-line-adopt first — it looks at your repo and tells you which…
★not rated 4
changed 14d agoA
tokens not measured
originalApache-2.0
Skills for code you'll have to live with. Two collections: craft (code that's understandable, named well, lean, loud about failure, and testable — without turning every edit into a rewrite) and rigor (claims you can believe — adversarial testing, full-suite banking, cited numbers, population sweeps, evidence-backed…
★not rated 4 4mo agoA
tokens not measured
originalMIT
Connects the glim.sh MCP server (live Twitter/X, Reddit, web, GitHub, Amazon, YouTube, Telegram data) and bundles a usage skill with tool prices and query patterns.
★not rated 4 1mo agoA
tokens not measured
originalMIT
A template for turning an exhibition plan into a three-page presentation outline. The pages cover the theme, its explanation and visitor story, and a detailed area-by-area display plan.
SEO data platform integration guide covering Google Search Console API, GA4 Data API, PageSpeed Insights API v5, CrUX API, and third-party data tools (Bright Data SERP, DataForSEO). Provides API usage patterns, data analysis methodologies, and reporting automation. Use when user says "Search Console", "GSC", "GA4"…
A decision framework for content operators: given a topic, decide whether to write it, and if so, with what priority. Distilled from real-world brand content operations.
A command for searching for jobs each day and tracking the application process. The input also refers to candidate details and companies already contacted.
Dokploy MCP server for Codex, Cursor, and Claude with compact stdio and hosted HTTP modes. Runs locally from the @vibetools/dokploy-mcp npm package. Needs 3 environment variables to run.
★not rated 4 3mo agoA
tokens not measured
originalMIT
MCP server for Tessera's LLM cost-optimization layer: anchored spend, drift, recommendations. Runs locally from the @tessera-llm/mcp-server npm package. Needs 2 environment variables to run.
★not rated 4 4mo agoA
tokens not measured
originalApache-2.0
Before you execute a drafted plan/spec/design doc, run a multi-agent adversarial cross-review to surface the flaws that cause downstream rework, aggregating "evidenced, cross-family-verified" objections into a verdict. Use when the user asks to "review this plan/spec", "can this approach be executed", "poke holes /…
AUTO-TRIGGER: Apply this skill when the user needs to defend, justify, or grow a marketing budget in front of a CEO, CFO, board, or executive team. Trigger phrases include: "budget review," "defending the budget," "CEO wants to cut marketing," "justify marketing spend," "prove ROI," "board presentation," "marketing…
Curious, non-committal mapping session modeled after how the human mind wanders through unfamiliar territory — building a mental model without reaching for a conclusion. Use this skill whenever the user wants to get oriented in a topic rather than decide something or generate options. Trigger on: "/explore", "help me…
Comprehensive guide for implementing Syncfusion Blazor dropdown components including AutoComplete, ComboBox, DropDown List, MultiSelect Dropdown, and ListBox. Use this when building selection interfaces, data binding, filtering, cascading dropdowns, custom templates, and accessible dropdown experiences in Blazor…
Skills for maintaining and developing Visual Basic 6 codebases with Claude Code.
★not rated 4 4mo agoA
tokens not measured
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: