public-safe

面向西综备考的 Codex Skill：经过真实学习场景校准，用于笔记审稿、错题复盘、学习状态判断与下一步训练建议。公开版不分发讲义、真题、答案或个人资料，只发布 public-safe 学习助教工作流。

Governance-bound Python security research runtime for controlled AI-assisted workflows, with policy gates, execution receipts, evidence traces, and operator-visible state; a precursor to the SCLite/GovEngine/RExecOp architecture.

Schema-backed contract layer for governed AI/security workflows: intent, policy decisions, approvals, execution receipts, evidence, and replayable audit traces.

Martin's Awesome Skills: public-safe Hermes and OpenClaw agent skills with Telegram approval workflows

Governance kernel for SCLite-backed workflows, validating policy gates, lifecycle boundaries, receipts, and trust decisions without owning execution.

Public-safe falsifier lab for observer-aware AI research: synthetic demos, evidence gates, negative controls and witness logs.

Public-safe UI boundary for local-first agent-state observability experiments; no private runtime connectors or diagnostic claims.

Public index for sanitized repositories, boundaries, whitepapers and support links.

MIT synthetic simulation sandbox for observable runs, falsifier examples and public-safe research boundary demos.

Public-safe OLL4 mirror of chopratejas/headroom with curated attribution and UI polish.

Public-safe electrical and controls architecture studies covering control narratives, sensor/actuator patterns, safety boundaries, and validation review.

Public derivative EPI scaffold for attention-head pruning research framing, review gates, and claim boundaries.

Public proof surface for simulation-first solar hydrogen control logic, safety gating, and ledger discipline.

Public mission-critical field-discipline notes for electrical, controls, and infrastructure review boundaries.

Public-safe standards-aware validation templates for safety checklists, test procedures, risk registers, and proof limits.

Public-safe PowerShell tool for auditing GitHub repository security posture, workflows, branch protection, GitHub Pages and static content findings across multiple repositories.

Public-safe data and model infrastructure studies covering source capture, metadata, vector indexing, dataset versioning, fine-tuning boundaries, and release cards.

Public-safe DGX/NVIDIA workstation documentation patterns without exposing topology, endpoints, credentials, or sealed IP.

Public-safe automated engineering system patterns covering memory, source control, workflow automation, run logs, mirrors, and review loops.

Local-first context mesh for agent workflows: a Rust CLI and MCP stdio runtime for durable events, rebuildable context graphs, skill manifests, public-safe exports, and agent mailboxes.