@danielfoojunwei

hot-swap-context builds durable, user-owned memory infrastructure for agents by scaffolding typed context vaults, governance policies, and portable bundles that survive runtime changes. three workflows (creation, audit, migration) with concrete deliverables and mcp-native interfaces.

auto-feedback analyzes execution logs, detects performance regressions, auto-generates regression tests, and produces signed improvement reports. operates standalone or integrates with intent-engineering and dark-factory for richer analysis across repeated cycles.

agentic-wireless-manager provides detailed steps for network scanning, signal scoring, and SAC-LTC inference across macOS, linux, and windows, but lacks concrete error handling, fallback logic, and validation steps between procedural stages.

agent-see converts websites and saas products into agent-executable bundles through a multi-stage pipeline. covers conversion, verification, launch artifacts, plugin packaging, and deployment orchestration with explicit output artifacts and status dashboards at each step.

intent-engineer describes a meta-framework for agent self-governance and multi-skill orchestration, but remains largely conceptual. the skill reads as architectural vision rather than actionable procedure, with minimal concrete trigger phrases and no documented failure modes.