Ruban S
Community@ruban-s · Coimbatore,TamilNadu,India
Ruban S maintains DevOS, a 38-skill developer harness spanning multi-agent research, adversarial reasoning, web scraping, evaluation frameworks, and TypeScript CLI scaffolding.
Agent Skills by Ruban S
Showing 38 vetted skills indexed across 1 GitHub repositories.
ruban-sResearch
Runs multi-agent web research with URL verification and confidence-tagged findings across four depth modes.
ruban-sRootCauseAnalysis
Investigate failures using Five Whys, Fishbone, Postmortem, Fault Tree, and Kepner-Tregoe methods.
ruban-sApertureOscillation
Analyzes one design question at component and system zoom levels to surface scope mismatches.
ruban-sCortex
Manages a typed knowledge archive of linked notes and recalls prior sessions and conversations.
ruban-sUpgrade
Extracts upgrade techniques from Anthropic sources and practitioner feeds into evidence-backed recommendations.
ruban-sEvals
Grades AI agent outputs with deterministic assertions and LLM judges across multi-trial eval suites.
ruban-sWorldThreatModel
Tests ideas and strategies against eleven horizon reads spanning 6 months to 50 years.
ruban-sOptimize
Runs autonomous hill-climbing loops that mutate code or prompts, measure outcomes, and keep only winning changes.
ruban-sCreateCLI
Generates production-ready TypeScript CLIs using a three-tier template system with Bun.
ruban-sISA
Drafts, audits, and merges Ideal State Artifact specification documents with verifiable claims.
ruban-sExtractWisdom
Extracts content-adaptive wisdom sections from videos, podcasts, and articles at five depth levels.
ruban-sCouncil
Convenes multi-agent debates with topic-briefed personas producing round-by-round transcripts and synthesis.
ruban-sPrompting
Generates and validates prompts programmatically using Handlebars templates, YAML data, and prompt engineering standards.
ruban-sBitterPillEngineering
Audits AI instruction files to cut redundant, contradictory, or vague rules and estimate token savings.
ruban-sRedTeam
Attacks plans and ideas with parallel expert critics and returns severity-ranked rebuttals.
ruban-sBrightData
Scrapes blocked web pages by escalating through fetch, curl, Chrome render, and Bright Data proxy tiers.
ruban-sCreateSkill
Scaffolds, validates, tests, and optimizes DevOS skill units end to end.
ruban-sScience
Applies the scientific method as a structured problem-solving loop with falsifiable hypotheses and measured experiments.
ruban-sTldraw
Create, read, and edit tldraw .tldr canvas files as validated JSON records.
ruban-sHTML
Renders session output into self-contained designed HTML artifacts via a deterministic TypeScript renderer.
ruban-sHardening
Strengthens existing tests and claims using property-based testing and mutation analysis.
ruban-sDevOS
Installs and maintains a verification-gated AI coding harness with spec interviews and health checks.
ruban-sMigrate
Classifies external content chunks against the DevOS taxonomy and commits them with provenance.
ruban-sSystemsThinking
Analyzes complex systems using causal loop diagrams, archetypes, and Meadows' leverage points.
Frequently Asked Questions About Ruban S
FAQPage SchemaWhat tasks can I accomplish with Ruban S's DevOS skill collection?▼
The 38 skills cover verified multi-agent web research, root cause analysis with Five Whys and FMEA, adversarial red-team critique, arXiv paper retrieval, web scraping via Bright Data and Apify actors, TypeScript CLI generation, tldraw canvas editing, HTML artifact rendering, and assertion-first evaluation of model outputs with pass@k scoring.
Who is the target user for these skills?▼
Developers and technical operators running an AI-assisted harness in their repositories. Skills like DevOS setup, CreateSkill, CreateCLI, Hardening, and Evals assume repo-level work with Read/Write/Bash access, while Research, Council, and RedTeam serve engineers needing structured investigation and adversarial review.
How do the evaluation and testing skills work in practice?▼
Evals grades output using deterministic asserts plus a structured judge over input-to-assert cases with pass^k and pass@k scoring for capability and regression suites. Hardening strengthens existing tests through property-based testing, mutation testing, and complexity analysis rather than writing new feature tests.
What are the platform and dependency requirements?▼
Several skills are platform-bound: CMUX and Interceptor are Mac-only, with Interceptor driving real Chrome or Brave via macOS Computer Use. BrightData escalates to metered Bright Data proxies, Apify requires hosted actors, and Fabric runs 240+ prompt patterns with CLI reserved for YouTube transcripts.
Are these skills open source and what do they cost?▼
The manifest does not state a global license; Webdesign references complete terms in LICENSE.txt, and Evals is described as subscription-billed. Bright Data proxy usage is metered, so scraping at the highest escalation rung incurs per-request cost.