Recently updated open-source projects

skill-company-research creates a structured company report for product managers, covering products, executive thinking, customers, pricing, market position, PM culture, and Hamilton Helmer’s Seven Powers framework. It can save product screenshots alongside the Markdown report, with modes for full research, strategy synthesis, competitive analysis, and interview preparation. The result is a reusable company reference rather than a one-question answer.

This skill audits a page for visibility in Google AI Overviews, ChatGPT Search, Perplexity, Gemini, and Copilot. It checks foundational SEO, debunked GEO tactics, extractability, authority, and citation readiness, then scores three pillars and proposes prioritized fixes; it reports what cannot be assessed without page access or performance data.

The Metric Mandate guides an AI project through five concrete decisions: the number that should change, its current baseline, the minimum worthwhile improvement, when it will be measured, and what outcome should stop the project. It rejects vague answers and stress-tests whether a metric can be improved without solving the real problem. The result is a concise table of success criteria and metric gameability before implementation begins.

decision-lens asks Claude to examine a complex choice from independent product, marketing, UX, business, risk, and other specialist perspectives. The result separates consensus from disagreement, gives one primary recommendation, and lists questions that only the decision owner can answer, followed by concrete next actions. It is meant for trade-off decisions, not coding tasks or simple fact questions.

This skill reads the latest 10-K beside an earnings-call transcript and writes an analyst memo with a brief, highlights, gap analysis, call-only color, and follow-up questions. Numbers are attributed to the filing or call, missing sections are flagged rather than guessed, and the result is a research aid rather than investment advice.

This skill fetches papers and, when available, arXiv TeX sources, then resolves BibTeX through open scholarly services. It requires each factual claim to cite a section or table/figure with a page and writes a missing marker instead of inventing a citation; the workflow is slower and more expensive than a casual summary.

This skill offers a pre-viewing guide with historical context, craft details, viewing cues, mindset, and version availability without spoilers. After the film, it switches to full-spoiler analysis of production history, formal choices, contemporary reception, and what a modern viewer may miss, selecting only the lenses that fit instead of producing a generic review.

This skill extracts a thesis structure and citation inventory, then assigns subsections to parallel reviewers who verify authors, years, pages, source identity, and text-to-bibliography matches against PDFs. It consolidates the results into a Markdown report with an executive summary and prioritized fixes; it does not edit the thesis or invent unavailable citations.

This Claude skill drafts or critiques X posts against a policy set derived from public ranking code, Community Notes materials, and cited analyses. It can run a pre-flight review, explain failed checks, suggest rewrites, and advise on format, timing, or media. Confidence labels distinguish confirmed evidence from inference, and the project explicitly does not guarantee reach because ranking systems and learned weights change.

Kavach Skill guides a coding agent through adding Kavach gates, policy loading, drift detectors, signed permit tokens, audit chains, and default-deny enforcement in Python or Node/TypeScript services. It also covers observe-only rollout, hot-reloading, and encrypted, signed, replay-protected channels, with the two SDKs backed by the same Rust core. The skill follows the published SDK surface; deeper HTTP or MCP middleware and multi-replica Redis deployments are outside its scope.

This Claude Code and Codex skill researches a topic through parallel sector teams, cross-sector pairings, and an independent verification pass against analyst, vendor, and trade coverage. It produces sector reports, opportunity and trend documents, a verification log, and a curated action shortlist, removing claims that are already widely priced in; scheduled recurring runs are supported.

Use this skill to create or update credentials in 1Password, generate commit-safe .env.tpl files containing op:// references, and inject resolved values only when a command runs. It also documents rotation and integrations for common development tools. The 1Password CLI v2 must be installed and signed in, and the recommended op run pattern avoids unsafe source or eval-based injection.

This skill reads the actual Claude usage reset time from the Anthropic usage API, adds a five-minute buffer, and schedules a deferred task for when capacity returns. It depends on Claude Code’s schedule skill and a macOS Keychain credential named Claude Code-credentials, so it is tied to that environment.

Ask about AI regulation, policy, law, or governance to receive a cited conversational answer, a topic briefing, a recent-development update, or a one-question-at-a-time quiz. The skill prioritizes an editable list of regulators, standards bodies, legal analysis, research, and trackers, and uses live retrieval for time-sensitive topics. Outputs can be reshaped for Slack, Confluence, email, or social posts.

This skill structures a product’s pre-launch legal and privacy review around source documents, user paths, checklists, prioritization, and a 5×5 inherent-versus-residual risk matrix. It produces a summary and detailed review with Tier 1–3 workstreams, and can optionally publish to Confluence; it is a working aid, not legal advice or sign-off.

This skill supplies prompt templates and a reference style for educational illustrations of quantum computers, dilution refrigerators, low-temperature experiments, and chip placement. Its visual language uses colored-pencil texture, an off-white paper background, pale blue and lavender accents, and generous whitespace; it works with Claude Skills or other image generators.

This terminal skill designs a course around the learner’s goal, baseline, pacing, and target ability, then teaches it across resumable sessions. Each unit uses a pre-test, productive-failure attempt, lesson, practice, project, checkpoint, and spaced-retrieval update; it stores the student model locally and requires plan approval before content generation.

performance-audit-skill audits pages through PageSpeed Insights, Lighthouse, and Core Web Vitals, then prioritizes fixes by measurable impact, implementation cost, and risk to layout, tracking, UTMs, forms, SEO, or accessibility. It supports auditing an existing URL or code before deployment and is available through adapters for many coding agents. Use it for performance regressions or speed-related conversion concerns.