Shape agent capabilities into your workflow.

Guard git clone, npm, pip, cargo, and similar install requests with a three-stage blocklist, trust check, and limited code sniff. Unknown sources receive a small inspection of install scripts or suspicious patterns, while blocked results include the threat and attribution. The skill runs in memory without network calls or executing the requested install command.

Goal by Goal turns a plan, PRD, or brief into 5–12 sequenced goals, with a brief, tracker, reviewer prompt, and check script for each. Build, test, or lint checks are required, and each goal must pass its read-only check before a commit.

Russian Humanizer rewrites machine-sounding Russian by removing bureaucratic phrasing, translationese, AI clichés, and mechanical typography while preserving meaning, terminology, genre, and voice. It applies only clear low-risk changes automatically and explains riskier suggestions with their rule source and exceptions. The project explicitly prioritizes authentic prose over attempts to evade AI detectors.

The builder interviews you about role, topics, channels, tone, and hard style rules, then analyzes 5–20 real samples or a Telegram export for recurring phrases, rhythm, and signature markers. It generates an invokable Claude Code skill plus a reusable voice profile that can be updated with new writing later. The analyzer uses only Python’s standard library, but sparse or overly polished samples produce a less faithful result.

Archetype Writer combines twelve Jungian voice personas with channel guides for websites, LinkedIn, email, UI, and short microcopy. It can cap reading difficulty from A2 to C2 and refine cadence, register, hooks, compression, and variation, working as a Claude Code or Claude.ai skill or as supplied context in ChatGPT.

TRS-300229 is a Chinese research skill for company 300229, covering company and industry context, peers and valuation, SaaS transition, workflows, source maps, query templates, checklists, and risk matrices. It asks users to record dates and links, separate facts from views, cross-check with two sources, keep financial definitions consistent, and avoid guessing.

Point tracedocs at a local repository or Git URL to create a study-docs package covering setup, operations, deployment, architecture, APIs, data models, troubleshooting, and maintenance. It also writes an index.json for coding agents. Important claims cite source files and carry Verified, Inferred, or Unknown labels, so undocumented deployment steps are recorded as gaps rather than guessed.

spring-boot-pr-review-skill reviews Java and Spring Boot changes for production failures that generic reviewers often miss. It looks for bypassed proxy annotations, silent transactional-event behavior, N+1 queries, tenant-unsafe cache keys, missing optimistic locking, and external calls held inside database transactions. Findings explain the underlying mechanism, assign a calibrated severity, and suggest a concrete fix rather than focusing on superficial style.

Skill ClaudeMD helps trim an existing CLAUDE.md by deciding what to keep, cut, reference, or move into path-scoped instructions, skills, nested files, and hooks. It checks commands, paths, and contradictions, reports before-and-after size, and targets a soft 200-line ceiling. It does not rewrite MEMORY.md or author a file from scratch.

This Bash tool diagnoses a missing Ask Gemini sidebar button caused by Chrome’s local Variations region seed changing to a non-US code. It backs up the Local State file, rewrites four country fields to `us`, and verifies both the JSON and the resulting values, with a restore option available. Chrome must be fully quit before the edit, and the setup also assumes current desktop Chrome, US English display language, and a US network route.

Source a session transcript, file, or topic, extract the concepts that actually matter, and turn them into a dated checklist. The skill asks one question at a time, probes why a decision was made, marks confirmed items immediately, and rephrases questions when needed. It does not offer to finish until the checklist is complete, with an optional mode for teaching another person.

Scaffold a runnable MapLibre web map from a description with standard coordinate systems, open basemaps, colorblind-safe palettes, attribution, and documented data provenance. The included Kenya demonstration map is illustrative: its town centroids and intensity labels are editorial, not exact event measurements. The skill also provides a checklist for auditing the result.

Polish Traditional Chinese writing for Taiwanese usage by removing formulaic openings, filler, translated syntax, and Mainland Chinese terminology. The skill combines language guidance with a terminology scanner, a six-dimension score, and a CJK-specific check for broken Markdown emphasis. It can be used while writing, editing, or reviewing Obsidian notes and other Chinese articles.

Build a Chokepoint Brief by anchoring demand, mapping a supply chain down to its physical dependencies, testing whether a bottleneck is truly irreplaceable, and listing catalysts and thesis-killers. The method can generalize beyond semiconductors, but it is an unofficial educational reconstruction from public material, not financial advice or a recommendation to buy named companies.

Create Vietnamese proposals, research reports, meeting minutes, or summary documents with prescribed structure, typography, margins, bullets, justification, and objective third-person language. The skill includes a validation checklist and a python-docx blueprint, but it is a formatting aid rather than a legal authority; local requirements should still be confirmed.

Give the skill a country and health domain to map health-system context, digital systems, indicators, gaps, and stakeholder questions from open institutional and academic sources. It renders a linked HTML profile and compact JSON, keeps contradictions visible, and records missing evidence instead of filling gaps. This is a starting brief, not a clinical tool or final assessment, and needs country-informed expert confirmation.

Manage each Slidev presentation as its own repository: scaffold it, define a palette, run review rounds, polish layout, build, and export PDF, PPTX, or PNG. Approval requires at least two review rounds, and the generated GitHub Actions workflow publishes to Pages on pushes. Setup needs GitHub CLI authentication, Node, and a configured destination directory.

Use one lead Claude Code session to spawn interactive workers in tmux, give each a task brief, coordinate through shared files, monitor progress, and require a different worker to challenge each result. The harness avoids servers and daemons but consumes parallel model usage and grants workers real filesystem and network permissions. Run it in a disposable or well-controlled environment.