yoonation/ai-grounded
Harness engineering for AI coding agents. Guides steer the agent before it acts, sensors catch what it got wrong after: a project constitution and concern catalogs, 12 read-only review agents, a 12-gate commit chain, and loop closure that b
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
Copy this rig
# review before running: this installs third-party code
$ npx degit yoonation/ai-grounded/.claude ./rig-ai-grounded # inspect, then merge into .claude/ MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ npx degit yoonation/ai-grounded/.claude/skills/commit-messages .claude/skills/commit-messages $ npx degit yoonation/ai-grounded/.claude/skills/security .claude/skills/security $ npx degit yoonation/ai-grounded/.claude/skills/speckit-analyze .claude/skills/speckit-analyze $ npx degit yoonation/ai-grounded/.claude/skills/speckit-checklist .claude/skills/speckit-checklist $ npx degit yoonation/ai-grounded/.claude/skills/speckit-clarify .claude/skills/speckit-clarify $ npx degit yoonation/ai-grounded/.claude/skills/speckit-constitution .claude/skills/speckit-constitution $ npx degit yoonation/ai-grounded/.claude/skills/speckit-git-commit .claude/skills/speckit-git-commit $ npx degit yoonation/ai-grounded/.claude/skills/speckit-git-feature .claude/skills/speckit-git-feature $ npx degit yoonation/ai-grounded/.claude/skills/speckit-git-initialize .claude/skills/speckit-git-initialize $ npx degit yoonation/ai-grounded/.claude/skills/speckit-git-remote .claude/skills/speckit-git-remote $ npx degit yoonation/ai-grounded/.claude/skills/speckit-git-validate .claude/skills/speckit-git-validate $ npx degit yoonation/ai-grounded/.claude/skills/speckit-implement .claude/skills/speckit-implement $ npx degit yoonation/ai-grounded/.claude/skills/speckit-plan .claude/skills/speckit-plan $ npx degit yoonation/ai-grounded/.claude/skills/speckit-specify .claude/skills/speckit-specify $ npx degit yoonation/ai-grounded/.claude/skills/speckit-tasks .claude/skills/speckit-tasks $ npx degit yoonation/ai-grounded/.claude/skills/speckit-taskstoissues .claude/skills/speckit-taskstoissues $ npx degit yoonation/ai-grounded/.claude/skills/speckit-workflow-post-impl .claude/skills/speckit-workflow-post-impl $ npx degit yoonation/ai-grounded/.claude/skills/speckit-workflow-post-plan .claude/skills/speckit-workflow-post-plan $ npx degit yoonation/ai-grounded/.claude/skills/speckit-workflow-post-spec .claude/skills/speckit-workflow-post-spec $ npx degit yoonation/ai-grounded/.claude/skills/speckit-workflow-pre-commit .claude/skills/speckit-workflow-pre-commit $ npx degit yoonation/ai-grounded/.claude/skills/terraform .claude/skills/terraform
$ curl -fsSL --create-dirs -o .claude/agents/adr-architect.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/adr-architect.md $ curl -fsSL --create-dirs -o .claude/agents/closure-auditor.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/closure-auditor.md $ curl -fsSL --create-dirs -o .claude/agents/code-reviewer.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/code-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/concern-selector.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/concern-selector.md $ curl -fsSL --create-dirs -o .claude/agents/discovery-agent.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/discovery-agent.md $ curl -fsSL --create-dirs -o .claude/agents/operational-architect.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/operational-architect.md $ curl -fsSL --create-dirs -o .claude/agents/performance-reviewer.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/performance-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/production-readiness.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/production-readiness.md $ curl -fsSL --create-dirs -o .claude/agents/security-reviewer.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/security-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/staff-engineer.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/staff-engineer.md $ curl -fsSL --create-dirs -o .claude/agents/test-architect.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/test-architect.md $ curl -fsSL --create-dirs -o .claude/agents/threat-modeler.md https://raw.githubusercontent.com/yoonation/ai-grounded/main/.claude/agents/threat-modeler.md
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"hooks": {
"PreToolUse": [
{
"matcher": "Edit|Write",
"hooks": [
{
"type": "command",
"command": "[ \"$(git branch --show-current)\" != \"main\" ] || { echo '{\"block\": true, \"message\": \"Cannot edit on main branch. Create a feature branch first.\"}' >&2; exit 2; }"
}
]
},
{
"matcher": "Bash",
"hooks": [
{
"type": "command",
"command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/block-dangerous-commands.sh"
}
]
},
{
"matcher": "Read|Edit|Write",
"hooks": [
{
"type": "command",
"command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/block-secret-file-access.sh"
}
]
}
]
}
} Skills (21)
commit-messagessecurityspeckit-analyzespeckit-checklistspeckit-clarifyspeckit-constitutionspeckit-git-commitspeckit-git-featurespeckit-git-initializespeckit-git-remotespeckit-git-validatespeckit-implementspeckit-planspeckit-specifyspeckit-tasksspeckit-taskstoissuesspeckit-workflow-post-implspeckit-workflow-post-planspeckit-workflow-post-specspeckit-workflow-pre-committerraform
Subagents (12)
| adr-architect model: sonnet | Drafts Architecture Decision Records (ADRs) from existing artifacts. Reads spec, plan, threat-model, clarify answers, and constitution to produce ADR drafts that pre-fill Context, Consequences, and Al |
| closure-auditor model: opus | Verifies that pending-resolution items from upstream agents (staff-engineer, threat-modeler, performance-reviewer, production-readiness, operational-architect, code-reviewer, security-reviewer) have v |
| code-reviewer model: sonnet | Reviews code for quality, correctness, maintainability, and adherence to canonical engineering practices (SOLID, GoF, Clean Code). Detects anti-patterns and Fowler's code smells. Produces findings wit |
| concern-selector model: sonnet | Produces the per-feature resolved routing plan (feature-concerns.yaml) once at post-spec, within the bounds of the project profile. Runs the profile dial for the deterministic rigor floor, then reads |
| discovery-agent model: haiku | Fresh-eyes reuse survey before construction. Given the generated capability index (the machine-built map of what the codebase already exports plus the libraries already in the manifest), surveys which |
| operational-architect model: opus | Designs operational artifacts for production services - SLI/SLO/error budget definitions, observability instrumentation plans, and runbook scaffolding. Applies when the project has user-facing service |
| performance-reviewer model: sonnet | Proactive resource-safety review. Identifies patterns likely to cause CPU exhaustion, memory leaks, runaway loops, unbounded cost, or scaling failures BEFORE they reach production. Aware of cloud-spec |
| production-readiness model: opus | Combined DevOps/Platform engineer agent. Reviews specs, plans, and code for production operability, environment topology awareness, deployment readiness, observability, resilience, and reliability. Co |
| security-reviewer model: opus | Post-implementation security review. Catches OWASP Top 10 vulnerabilities and AI-specific threats. Verifies code-level mitigations for must-mitigate threats from threat-modeler. Enforces audit-envelop |
| staff-engineer model: opus | Provides principled engineering pushback on specs, plans, and major decisions. Apply YAGNI, DRY (with judgment), scope discipline, timing analysis, assumption challenging, language/framework/library s |
| test-architect model: opus | Designs tests from the spec's perspective, not the implementation's. Tests business use cases, not code paths. Coordinates with performance-reviewer to produce performance test cases. Flags over-mocki |
| threat-modeler model: opus | Proactive security analysis against specs and designs before implementation. Walks STRIDE, OWASP LLM Top 10, OWASP Agentic ASI Top 10, and MITRE ATLAS catalogs. Surfaces compliance implications (GDPR, |
Hooks (4)
| event | matcher | runs |
|---|---|---|
| PreToolUse | Edit|Write | [ "$(git branch --show-current)" != "main" ] || { echo '{"block": true, "message": "Cannot edit on main branch. Create a feature branch first."}' >&2; exit 2; } |
| PreToolUse | Bash | "$CLAUDE_PROJECT_DIR"/.claude/hooks/block-dangerous-commands.sh |
| PreToolUse | Read|Edit|Write | "$CLAUDE_PROJECT_DIR"/.claude/hooks/block-secret-file-access.sh |
| PostToolUse | Edit|Write | "$CLAUDE_PROJECT_DIR"/.claude/hooks/lint-on-edit.sh |
Slash commands (2)
/review/security-check
Similar rigs
mahendrjy/dotfiles
My automate dotfiles
Skill Collector 535 tok ·
D4lD4r/inubit-mcp-server
Local stdio MCP server for operating and diagnosing INUBIT 8.1 integration servers from AI assistants such as Claude Code
Skill Collector 530 tok ·
alfredoperez/speckit-companion
VS Code extension for spec-driven development — manage specs, workflows, and steering docs for AI CLI tools (Claude Code, Gemini CLI, Copilot CLI)
Skill Collector 3.1k tok ·
CodyAMaughan/agent-first-engineering
A curriculum + an agent-agnostic scaffolder for agent-first software engineering: design your codebase so AI coding agents (Claude Code, Codex, Cursor) succeed.
Skill Collector 3.1k tok ·