All authors
Claude Skills by shear559
github.com/shear55915 skills0 installs18 views
- Fabius Archivumfabius's memory layer: maintain and retrieve project memory in the workspace's declared record store. Read a project's record before work; preserve decisions, provenance, history; gated writes. Ingests meetings, videos, books.Votes: 0GitHub stars: 2
- Fabius Catenafabius's on-chain layer: write and review EVM and Solana programs, transactions, on-chain reads, money-safe; SEAL artifacts with content-bound hash, signature, timestamp, pending or confirmed. Defensive only: hardens and proves, never weaponizes.Votes: 0GitHub stars: 2
- Fabius Cohorsfabius's agent-engineering layer: define and orchestrate agents, tool-using assistants, multi-agent systems: schema, permissions, single-vs-multi decision, orchestration patterns. Deterministic service wiring (n8n/Zapier-class) is fabius-machina.Votes: 0GitHub stars: 2
- Fabius Conciliumfabius's cross-model council: several models on ONE question: first opinions, blind peer ranking, synthesis. Costs up to 3N+1 calls, earned only when one strong model keeps failing the SAME way. Splitting WORK across agents is fabius-cohors.Votes: 0GitHub stars: 2
- Fabius Decorfabius's ship-grade design layer: build or review any UI, email, slide, chart, diagram, data visualization; generate images; responsive layout, focus states, contrast, ARIA-only a11y, RTL. Production-quality polish; copy is fabius-mercatus.Votes: 0GitHub stars: 2
- Fabius Disciplinafabius's engineering process: architecture, build, refactor, migrate; debug what is broken, flaky or CI-only, root-cause before any change; verify by evidence. UI craft is fabius-decor; agents are fabius-cohors; restraint briefs are fabius-parcus.Votes: 0GitHub stars: 2
- Fabius Doctrinafabius's ML/LLM-engineering layer: train, fine-tune, evaluate, serve and monitor models; inference endpoints, quantization, MLOps tracking, eval gates. Owns the MODEL an agent calls, not the agent (fabius-cohors); lab science is fabius-scientia.Votes: 0GitHub stars: 2
- Fabius Fortunafabius's markets and economics layer: equity, market and macro analysis, valuation, honest backtesting, portfolio construction, risk-first sizing. Sourced, risk-bounded, falsifiable: analysis, never a prediction, never personalized advice.Votes: 0GitHub stars: 2
- Fabius Ludusfabius's game-craft layer: small games that are fun and ship. Core loop first, juice, explicit state machine, pixel-art lane, one balance knob at a time, jam-sized scope. For any game, mechanic, playable, toy or win/lose/score state.Votes: 0GitHub stars: 2
- Fabius Machinafabius's automation layer: wire deterministic service-to-service workflows (n8n-class) and prove the wiring before it runs live: triggers, webhooks, crons, API glue. Fixed-behavior steps route here; generative LLM agents are fabius-cohors.Votes: 0GitHub stars: 2
- Fabius Mercatusfabius's go-to-market layer: positioning, one-action funnels, copy that converts for landing pages, launches, ads, emails, posts, outreach, pricing; also SEO and AI-answer visibility. Owns the message; the visual build is fabius-decor.Votes: 0GitHub stars: 2
- Fabius Parcusfabius's always-on lean core: say less, build less, change less, assume less. Restraint briefs (no over-engineering, no future-proofing, YAGNI) route here, even on a refactor. ALWAYS-ON: sits under whatever task layer is active, never instead of one.Votes: 0GitHub stars: 2
- Fabius Praesidiumfabius's defensive-security layer: threat-model first, OWASP pass, least-privilege hygiene, supply chain; every finding ships with severity, fix, test; for auth, input, payments, uploads, deps. Defensive only: hardens, never weaponizes.Votes: 0GitHub stars: 2
- Fabius Scientiafabius's scientific-research layer: empirical method for biology and data-heavy sciences: falsifiable hypotheses, experiment design, database-grounded facts, bioinformatics, chemistry and omics pipelines. Model training is fabius-doctrina.Votes: 0GitHub stars: 2
- Fabiusfabius router: load at the start of any non-trivial task; sets HOW to work before any specialist fires. One set of operating rules above every model: scout wide, strike narrow, route to the layer that owns the concern. Use on "fabius" or for end-to-end work.Votes: 0GitHub stars: 2