Find the open-source solution that fits.

Use dedicated modes to decompose automotive requirements, check INCOSE and EARS quality, review C code against MISRA-C, or examine ADAS levels, ASIL, AUTOSAR, and SOTIF concerns. The skill works in several AI tools that accept system instructions and is intended for early analysis, training, and prototyping. It is not a certified tool and cannot replace formal static analysis, traceability systems, or independent safety assessment.

Token-Diet lays out an ordered optimization workflow for LLM applications: prune retrieved context first, reuse cached material, route work by complexity, keep structured state, and compress prose only at the final step. It also defines metrics and guardrails for token use, latency, cost, and quality, while excluding code, schemas, legal text, URLs, and identifiers from linguistic compression.

Ask Claude for a 3D wireframe HTML presentation and the skill produces a single file that opens directly in a modern browser. Slides use Three.js backgrounds, horizontal snap scrolling, parallax depth, responsive layouts, and automatic theme-color transitions. The presentation depends on an internet connection because its external libraries are loaded from CDNs.

Choose a fictional character and source work, then guide the skill through source research, trait extraction, and construction of an activatable character profile in Claude Code. It captures core conflicts, behavior patterns, relationship dynamics, speech style, and performance boundaries, followed by dialogue, scenario, and boundary tests to validate the result.

Import chat records to score six dimensions including initiative, response quality, uniqueness, emotional disclosure, behavioral signals, and future-oriented language. The app visualizes relationship trends, highlights notable lines, and flags possible over-interpretation or one-sided interest. Records are parsed locally, but a text summary is sent to the DeepSeek API for analysis; the result is intended as a personal reference.

Give Claude Code a firm but non-abusive review persona that challenges premature architecture, unjustified rewrites, and other red flags before helping. Nonna can flag technical debt, missing tests, and oversized files, respond in Italian or English, and switch between Full, Lite, and Del Sud modes. Hooks and slash commands are Claude Code-specific; the skill itself is offline-first with no external runtime dependencies.

Opus Compatibility Scanner checks instruction files, subagents, skills, settings, hooks, manifests, and Anthropic SDK call sites for known differences between Opus 4.6 and 4.7. You can run a dual-version compatibility scan, a 4.7-only scan, or symptom-led diagnosis. Findings include severity and source citations, while reports and fixes remain opt-in and no files are changed by default.

Use this skill in Claude Code, Gemini CLI, or Codex to classify AI-system risk, map obligations, and prepare materials such as conformity assessments, Annex IV documentation, FRIAs, and post-market plans. Its guidance links work items to EU AI Act provisions and ISO/IEC 42001 and 27090 controls. It is decision support rather than legal advice or a substitute for notified-body assessment and certification.

Explore an application once through Playwright MCP to generate maintainable spec files, then replay them with the Playwright CLI instead of repeating browser-driven reasoning. The toolkit adds accessibility, console, network, and regression checks, fingerprints bugs across runs, and keeps tests and reports in the repository. It does not include managed cloud execution or human triage.

Snapshot every accessible Azure subscription through the signed-in az CLI, then evaluate cost waste against versioned Microsoft and FinOps Foundation rules. The tool produces a forwardable Markdown report and machine-readable findings with severity, confidence, and estimated GBP savings ranges. Its command wrapper blocks mutating Azure verbs, so it investigates opportunities but does not remediate them.

Use the skill to scaffold Joomla components, modules, plugins, or templates and follow current patterns for MVC, PSR-4 namespaces, dependency injection, service providers, manifests, migrations, and web assets. Its on-demand references also cover testing, packaging, update servers, and common Joomla 5-to-6 pitfalls; the documented target requires Joomla 5 or 6 with PHP 8.3 or newer.

watch-video uses yt-dlp to fetch videos or accept local files, takes captions when available, falls back to local whisper.cpp transcription, and extracts adaptive video frames with ffmpeg. The agent receives metadata, a transcript, and a frame index so it can inspect only the visuals needed for an answer, with fast, balanced, and accurate modes controlling cost. No API key or third-party transcription service is required, but local tools and an optional Whisper model are needed; Claude Desktop is not supported.

SceneLens prepares public video URLs or local files for Claude by extracting frames at scene changes, running OCR on each frame, and attaching a timestamped transcript. It first tries native captions, then falls back to Groq or OpenAI Whisper and automatically splits audio that exceeds the API size limit before merging timestamps. Accuracy is best for videos under ten minutes, output is capped at 100 frames, and authenticated private platforms are not supported.

crawl-budget-manipulation helps SEO teams plan experiments around low-competition, long-tail queries. It combines keyword grouping, repeatable content briefs, looped internal links, a publishing schedule, Search Console actions, and follow-up metrics for crawl and ranking changes. The method is aimed at new or moderately authoritative domains and explicitly excludes high-competition head terms; its outcome figures are reported expectations, not guarantees.

skill-validator audits a Claude Code SKILL.md for structural problems, missing fields, weak trigger descriptions, unreplaced placeholders, and marketplace readiness. It produces a 0–10 score, a separate triggering score, critical issues, warnings, strengths, and one concrete recommended next step for each problem.

description-optimizer evaluates a skill's current description across five dimensions, explains its weakest point, and creates three scored alternatives: Comprehensive, Tight, and Discovery-optimized. It compares the variants with the original and gives a clear recommendation, with YAML-ready text for direct reuse.

Describe a project in plain language and `/scaffold` routes the request to a stack-specific wizard, shows the complete plan, and waits for confirmation before generating code, configuration, and project guidance. `/bootstrap` surfaces relevant community skills, while `/sync` exports project memory to Git for reuse across machines and several coding tools. The collection includes more than 30 skills, although many topic-specific reference skills are still clearly marked stubs.

Start with a few questions about priorities, voice, audiences, and campaigns, then crawl the company site and research competitors and public reputation signals. The skill maps the findings to HubSpot's Business, Customers, and Brand Kit fields, respecting form constraints, and produces a tabbed HTML paste sheet with individually copyable values. It can also export product research, ICPs, and personas for future updates.