BigHugger
GH Repository · GanyuanRan

Aegis

Make AI coding agents architecture-aware: baseline-first, evidence-verified, drift-checked, and safe across long tasks.

stars
1,206
30-day movement
+103/day
Related entries
82
Connections
4
dsh-pluginPythonspec-kitspec-driven-developmentagent-skillsawsome-coding-pluginai-agentscodexarchitecture-driven-developmentopencodeclaude-codeai-codingsoftware-architecturecoding-agentsskill-packdiveevidence-drivennodebaseline-firstfirst-principles

Aegis is a skill pack that makes AI coding agents architecture-aware, described as baseline-first, evidence-verified, and drift-checked across long tasks. It ships configuration files for multiple agent tools: AGENTS.md, CLAUDE.md, and cursor rules.

Use it when you want coding agents like Claude Code, Codex, or OpenCode to follow an architecture baseline and verify changes against evidence instead of drifting during long tasks.

Use it to

  • Install architecture-aware skills into Claude Code via CLAUDE.md
  • Add baseline-first rules to Cursor projects
  • Configure AGENTS.md conventions for Codex or OpenCode
  • Keep long agent coding tasks aligned with a software architecture baseline
  • Apply spec-driven, evidence-checked workflows to coding agents

For Developers using AI coding agents on structured codebases

Role
skill-pack
Language
Python
Licence
MIT
Forks
52
Last push
2026-09-16
Latest release
v1.0.6 · 2026-05-05
Skills shipped
22
topicsai-coding-agentsarchitecture-driven-developmentagent-skillsclaude-codespec-driven-developmentcursor-rules