@raishin/vanguard-frontier-agentic

v3.12.1

Published

3 days ago

Cloud and zero-trust agentic workflow marketplace for skills, agents, rules, MCP references, and compliance-aware architecture.

0High
0Medium
0Low

alphazero

agentic cloud zero-trust compliance soc2 pci-dss nis2 nist mcp codex terraform

Vanguard Frontier Agentic

This is the edge of agentic intelligence — an enterprise-grade ecosystem for running AI agents at scale in environments where a wrong move is a board-level incident. It collects reusable skills, agents, rules, MCP references, and supporting assets for AWS, Azure, OCI, GCP, Alibaba Cloud, Huawei Cloud, Kubernetes, and Terraform — plus cross-functional Legal, HR, Marketing, Salesforce, .NET, Java, Kotlin, PHP, Python, TypeScript, FinOps, Accounting, Finance, Oracle NetSuite ERP, SAP S/4HANA + BTP ERP, Microsoft 365 / Dynamics 365, a cloud-neutral Databricks lakehouse/AI board, and a cloud-neutral Snowflake data-platform board agentic ecosystems.

This is not just cloud infrastructure tooling. It is agentic coordination: maestro routing, escalation-aware protocol, structured handoff between specialists, and refusal-by-default safety on every irreversible action. Cloud is one domain it operates in. Coordination, governance, and escalation are the product.

📊 Catalog at a glance

| Catalog | Count | | --- | --- | | Skills | 761 | | Agents | 735 | | Providers | 45 | | Install roles | 76 | | Rules | 1 | | MCP references | 3 |

🧠 Skills = step-by-step workflows an AI assistant can follow.
🤖 Agents = reusable expert roles for review, architecture, and operations.
📏 Rules = durable instructions for a specific AI harness.
🔌 MCP references = trusted notes for connecting tools to real systems.
🗂️ Catalogs = machine-readable indexes so tools can discover everything.

Works with: Claude Code · Codex · GitHub Copilot · Cursor · Gemini CLI · Kiro · and any other coding agent.

📦 Available on npm: @raishin/vanguard-frontier-agentic is published on the public npm registry.

[!IMPORTANT] 🔁 The GitHub owner changed — the npm package name did not. Both are true at once: this repository moved from Raishin/vanguard-frontier-agentic to VincentChuWaiChow/vanguard-frontier-agentic when the account was renamed, but the npm scope did not move with it.
| What changed | What did NOT change | | --- | --- | | GitHub owner → VincentChuWaiChow | npm package name → still @raishin/vanguard-frontier-agentic |
GitHub redirects the old URLs, so existing clones and links keep working, but update your git remote and any pinned marketplace source when convenient.
The npm package is still @raishin/vanguard-frontier-agentic — npm install @raishin/vanguard-frontier-agentic@latest is unchanged, and no republish under a new name is planned. An npm scope is a separate namespace from a GitHub account: renaming the account does not rename the scope, and renaming the scope would break every existing install. Treat @raishin/… in install commands, .npmrc registry config, and packed tarball names as correct, not as a leftover.

⚠️ ALPHA FINOPS BUNDLE: As of v1.8.0, this package includes 4 new experimental FinOps agents and 7 skills for cloud cost optimization, AI economics modeling, Kubernetes rightsizing, and FOCUS-spec normalization. All are marked lifecycle: experimental. See the board readiness memo for known limitations, risk mitigation, and 30-day diligence closure requirements. Use at your own risk in pre-production environments. Production deployment requires signed design-partner SOWs, Big 4 accounting validation, and SOC 2 Type II observation (≥150 days).

🛰️ Why Vanguard Frontier?

"Vanguard frontier" is not branding — it is an operating posture. This ecosystem is built for the front line of agentic deployment, where AI agents touch real production systems, real regulated data, and real legal exposure.

🏛️ Built for Fortune 50 / high-stakes environments. Every agent assumes the blast radius is enterprise-scale: regulated data, audited controls, and decisions that survive legal discovery. Refusal-by-default beats a fast path to a board-level incident.
⚖️ The Legal + HR ecosystem is proof of cross-functional agentic coordination. 28 specialist agents (Legal maestro + 12 specialists, HR maestro + 14 specialists) and 3 cross-functional protocol skills demonstrate that agents can hand off, escalate, and coordinate across organizational boundaries — not just answer in isolation.
🧾 Audit-ready, privacy-preserving, escalation-aware by design. Every review and live-guard agent emits a structured verdict (verdict, evidence_level, blockers, safe_next_actions, open_questions) that maps directly to SOC 2, PCI DSS, NIS2, NIST CSF, and ISO 27001 — no post-processing.
🛡️ Battle-tested against real compliance, governance, and risk workflows. These patterns are exercised against live IAM mutations, KMS destruction, litigation holds, RIF planning, and privacy reviews — the workflows where a generic agent gets an organization sued.

The bar: an auditor, a regulator, or opposing counsel should be able to read the agent's output and trace exactly who decided what, on what evidence, and who approved the risk.

🧱 What's Inside — the three-layer agentic architecture

Vanguard Frontier is not a flat bag of prompts. It is a deliberate three-layer system, and every domain — cloud providers, Kubernetes, marketing, Legal, HR — follows the same shape.

| Layer | Role | Examples | | ----- | ---- | -------- | | 1. 🧭 Maestro (router) | Entry point. Classifies the request, routes to the right specialist, never executes risk itself. | legal-maestro-agent, hr-maestro-agent, kubernetes-maestro-agent, provider maestros | | 2. 🤖 Specialists | Domain experts with judgment and a hardened permission model. Each loads one companion skill and emits a guarded verdict. | 13 Legal specialists, 15 HR specialists, cloud advisory + live-guard operators | | 3. 🔗 Cross-functional protocol | Shared contracts that let specialists hand off, escalate, and coordinate across organizational boundaries without leaking scope. | legal-hr-routing-protocol, legal-hr-case-capsule, legal-hr-risk-taxonomy |

How it flows: a request enters at the maestro, which routes to a specialist. When a matter crosses a boundary — an HR investigation that needs privileged Legal review, or a Legal hold that triggers an HR data freeze — the cross-functional protocol carries a structured case capsule between agents, preserving privilege, minimizing data, and recording the escalation path.

This is what "agentic coordination" means here: routing, protocol, and escalation are first-class, not improvised.

🚀 Get Started

Pick the install path for your coding agent. Each dropdown is crystal-clear, step-by-step, and one-click plug-and-play where the harness supports it; the npm/export path works for everything else.

At a glance — which path is yours:

| Your harness | Fastest path | One-liner | | ------------ | ------------ | --------- | | 🤖 Claude Code | Plugin marketplace | /plugin marketplace add VincentChuWaiChow/vanguard-frontier-agentic | | 🐙 GitHub Copilot CLI | Plugin marketplace | copilot plugin marketplace add VincentChuWaiChow/vanguard-frontier-agentic | | 🖱️ Cursor | Clone + register plugin dir | git clone … then Settings → Plugins → Add Plugin Directory | | ⚡ Codex | Plugin marketplace | codex plugin marketplace add VincentChuWaiChow/vanguard-frontier-agentic | | ♊ Gemini / Antigravity | npm export | npx vfa-export-agents --platform gemini --all --repo . | | 🔮 Kiro | Add Powers per-directory | Powers panel → Add Custom Power → Local Directory | | 📦 Any other | npm + vfa-export-agents CLI | npm install @raishin/vanguard-frontier-agentic@latest |

Expand the matching dropdown below for the full step-by-step.

/plugin marketplace add VincentChuWaiChow/vanguard-frontier-agentic
/plugin install vanguard-frontier-agentic@vanguard-frontier-agentic

Or wire it into ~/.claude/settings.json (or your project's .claude/settings.json) for team-wide trust:

{
  "extraKnownMarketplaces": {
    "vanguard-frontier-agentic": {
      "source": { "source": "github", "repo": "VincentChuWaiChow/vanguard-frontier-agentic" }
    }
  },
  "enabledPlugins": {
    "vanguard-frontier-agentic@vanguard-frontier-agentic": true
  }
}

Pin to a tag for reproducible installs: pick any released tag for stable versions, or use @latest for the current release.

Bundled: all 735 cloud, security, compliance, Kubernetes, Terraform agents (incl. provider maestros and live-guard agents)
Spec: .claude-plugin/marketplace.json + .claude-plugin/plugin.json (canonical Claude Code plugin layout)
Not bundled: skills, rules, MCP references — use the npm path for those
Docs: code.claude.com/docs/en/plugin-marketplaces

# Add the marketplace, then install
copilot plugin marketplace add VincentChuWaiChow/vanguard-frontier-agentic
/plugin install vanguard-frontier-agentic

Or in .github/copilot/settings.json for repo-wide trust:

{
  "extraKnownMarketplaces": [
    "https://raw.githubusercontent.com/VincentChuWaiChow/vanguard-frontier-agentic/master/.github/plugin/marketplace.json"
  ]
}

Marketplace manifest: .github/plugin/marketplace.json declares this repo as a single-plugin marketplace
Source path: ./ (the repo root is the plugin root)
Bundled: 735 Copilot agent adapters under agents/<provider>/<agent>/harnesses/copilot.agent.md
Docs: github.com/github/copilot-cli (/plugin marketplace add)

# Clone the repo, then register it as a plugin directory in Cursor:
git clone https://github.com/VincentChuWaiChow/vanguard-frontier-agentic

In Cursor: Settings → Plugins → Add Plugin Directory → pick the cloned repo path. Or via the Cursor Extension API:

vscode.cursor.plugins.registerPath("/absolute/path/to/vanguard-frontier-agentic");

Plugin manifest: .cursor-plugin/plugin.json enumerates all 735 Cursor agent adapters explicitly via the agents field
Bundled: all agents from agents/<provider>/<agent>/harnesses/cursor.agent.md
Rules: existing rules/ directory at repo root is auto-discovered by Cursor
Docs: cursor.com/docs/plugins · cursor.com/docs/reference/plugins

Kiro Powers UI is per-Power directory add — there is no single-command marketplace flow. This repo ships 45 Powers under powers/, one per provider, so Kiro users can add only what they need.

# 1. Clone this repo
git clone https://github.com/VincentChuWaiChow/vanguard-frontier-agentic
cd vanguard-frontier-agentic

# 2. In Kiro:
#    Open the Powers panel → "Add Custom Power" → "Local Directory"
#    Paste the absolute path to the Power(s) you need, one at a time:
#       /absolute/path/to/vanguard-frontier-agentic/powers/vanguard-aws
#       /absolute/path/to/vanguard-frontier-agentic/powers/vanguard-kubernetes
#       /absolute/path/to/vanguard-frontier-agentic/powers/vanguard-terraform

Powers available: vanguard-accounting, vanguard-alibaba, vanguard-argocd, vanguard-aws, vanguard-azure, vanguard-backstage, vanguard-cert-manager, vanguard-cilium, vanguard-contabo, vanguard-databricks, vanguard-dotnet, vanguard-falco, vanguard-finance, vanguard-fluxcd, vanguard-frontend, vanguard-gcp, vanguard-generic, vanguard-hetzner, vanguard-hr, vanguard-huawei, vanguard-ionos, vanguard-istio, vanguard-java, vanguard-kotlin, vanguard-kubernetes, vanguard-kyverno, vanguard-legal, vanguard-marketing, vanguard-microsoft, vanguard-multi-cloud, vanguard-netsuite, vanguard-nvidia, vanguard-oci, vanguard-opentelemetry, vanguard-ovhcloud, vanguard-php, vanguard-prometheus, vanguard-python, vanguard-salesforce, vanguard-sap, vanguard-scaleway, vanguard-sigstore, vanguard-snowflake, vanguard-terraform, vanguard-typescript
Each Power ships: routing pattern (maestro entry), live-mutation discipline, provider invariants (account-ID/region, MLPS 2.0, EU sovereignty, etc.)
Frontmatter: strict-5 fields (name, displayName, description, keywords, author) per Kiro spec
For Kiro agent adapter files (.kiro/agents/*.md, .kiro/agents/*.json): use the npm-export path below
Docs: github.com/kirodotdev/powers

Antigravity reads skills from .agent/skills/<name>/SKILL.md (workspace) or ~/.gemini/antigravity/skills/<name>/ (global). There is no first-party marketplace install command — use the npm export to write skills + adapters into the right paths:

# Install the package
npm install @raishin/vanguard-frontier-agentic@latest

# Export agents + companion skills for Gemini Antigravity
npx vfa-export-agents --platform gemini --all --repo .

Or for a single provider:

npx vfa-export-agents --platform gemini --provider aws --repo .

Workspace skills install to: .agent/skills/<name>/SKILL.md
Global rules: ~/.gemini/GEMINI.md
MCP servers: configure via Antigravity's MCP Store UI → mcp_config.json
Docs: antigravity.google · github.com/google-gemini/gemini-cli

# Add the marketplace, then enable the bundled plugin
codex plugin marketplace add VincentChuWaiChow/vanguard-frontier-agentic
/plugin install vanguard-frontier-agentic@vanguard-frontier-agentic

codex plugin marketplace add writes the marketplace into your ~/.codex/config.toml. The resulting block looks like this (the screenshot pattern):

[marketplaces.vanguard-frontier-agentic]
last_updated = "2026-05-11T06:46:00Z"
last_revision = "<sha>"
source_type = "git"
source = "https://github.com/VincentChuWaiChow/vanguard-frontier-agentic.git"

[plugins."vanguard-frontier-agentic@vanguard-frontier-agentic"]
enabled = true

Marketplace registry: .agents/plugins/marketplace.json at repo root (canonical Codex location per codex-rs plugin-json-spec)
Bundled plugins:
- vanguard-frontier-agentic — the main plugin, manifest at plugins/vanguard-frontier-agentic/.codex-plugin/plugin.json
- cross-platform-agent-template — scaffold for new cross-platform agents
For agent adapter files (.codex/agents/*.toml): after enabling the plugin, run npx vfa-export-agents --platform codex --all --repo . to write the 735 agent adapters into your repo
Other commands: codex plugin marketplace upgrade vanguard-frontier-agentic, codex plugin marketplace remove vanguard-frontier-agentic
Docs: github.com/openai/codex · Codex plugin spec

Prerequisite: Node.js 18+

# 1️⃣ Install the package
npm install @raishin/vanguard-frontier-agentic@latest

# 2️⃣ Export agents for your role into your repo (claude-code shown — swap platform)
npx vfa-export-agents --platform claude-code --role cloud-security-engineer --repo .

# 3️⃣ Open your coding agent and reference the exported agent
#    "Use kubernetes-rbac-review-agent to audit this RBAC change."

Supports --platform: claude-code, codex, copilot, cursor, gemini, kiro, kiro-ide, kiro-cli. Supports --role, --agents, --all, --provider filters. See the Install Reference for the full argument matrix.

Install paths

There are now eight supported install paths — Claude Code plugin marketplace, GitHub Copilot CLI marketplace, Cursor plugin, Codex plugin marketplace, Kiro Powers, Gemini Antigravity skills, npm package + vfa-export-agents CLI, and the third-party skills CLI — each with different versioning, trust, and scope characteristics. See docs/integrations/skills-cli.md for the full trust matrix, verified flag syntax, pinning guidance, and pre-install inspection steps.

npm install @raishin/vanguard-frontier-agentic@latest

🖥️ Terminal UI (vfa-tui)

⚠️ Alpha — the TUI is functional but under active development.

An enterprise-grade terminal interface for interactive catalog browsing, validation gate execution, and export command building — built in Rust for speed and security by construction. Distributed as the vfa-tui crate.

Install (recommended — from crates.io):

cargo install vfa-tui

Or download a prebuilt binary (Linux / macOS / Windows, x86-64 + arm64) — each release attaches the binaries alongside their SBOM and checksums.sha256 — from the latest vfa-tui release.

Build from source (alternative):

cd tools/vfa-tui && cargo build --release

Run (from the repo root — auto-detects workspace):

vfa-tui                                      # installed via cargo or prebuilt binary
./tools/vfa-tui/target/release/vfa-tui       # from a source build

Key features:

🗂️ Catalog browsing with fuzzy search across all agents, skills, and providers
✅ Validation gate execution with real-time progress and structured output
📦 Export command builder with platform/role/provider selection
🔐 Security by construction — no shell injection, no credential handling, no network access

Full documentation

🧠 Skills

761 skills across AWS, Azure, OCI, GCP, Alibaba Cloud, Huawei Cloud, Kubernetes, CNCF ecosystem, Terraform, marketing governance, and more.

| Domain | ------------------ | 🟧 AWS | 🟥 OCI | 🟩 GCP | 🟦 Azure | 🟠 Alibaba Cloud | 🔴 Huawei Cloud | ☸️ Kubernetes | 🛡️ Kyverno | 🔄 Argo CD | 🕸️ Istio | 🐝 Cilium | 📡 OpenTelemetry | | 🟩 Terraform | 📣 Marketing | ☁️ Salesforce | Ⓜ️ Microsoft 365 / D365 | | 🧱 Databricks | ❄️ Snowflake | Count | What they cover | | ----: | ------------------------------------------------------------------------------------------------- | | 47 | IAM, EKS, ECS, Lambda, RDS, S3, Cost, DevOps, Bedrock, Security, WAF reviews, Live Guards | | 41 | ADB, OKE, IAM, Vault, Resource Manager, Cost, Networking, WAF reviews, Live Guards | | 51 | GKE, BigQuery, Vertex AI, Cloud Run, AlloyDB, Firebase, Gemini API, WAF reviews, Live Guards | | 36 | AKS, App Service, ARM/Bicep, Key Vault, PIM, Cost, Entra ID, CosmosDB, WAF reviews, Live Guards | | 43 | ACK, ECS, PolarDB, MaxCompute, RAM, OSS, MLPS 2.0, WAF reviews, Live Guards | | 43 | CCE, GaussDB, ModelArts, DEW, SecMaster, OBS, MLPS 2.0, WAF reviews, Live Guards | | 10 | RBAC review, workload identity, PSA, live RBAC/admission/mesh/network/ArgoCD guards, maestro | | 1 | ClusterPolicy/Policy, PolicyException, failureAction, background scan | | 2 | AppProject blast-radius, sync impersonation, RollingSync, sync-window | | 8 | Ambient mesh, authorization, traffic resilience, Gateway API, upgrades, diagnostics, routing, guarded live policy changes | | 1 | CiliumNetworkPolicy, ClusterMesh trust, 169.254.169.254 egress, WireGuard encryption | 1 | Collector pipeline, memory_limiter, receiver exposure, exporter cardinality, credential handling | | 10 | Engine-shared Terraform/OpenTofu board: maestro routing, plan blast radius, state reliability, estate reconciliation, supply-chain integrity, engine compatibility, policy evidence, execution governance, module contract, verification strategy | | 14 | Consent, pixel-leakage, martech access, GPC, email auth, ads.txt, targeting fairness, EU AI Act, audience uploads, list retention, influencer, dark patterns, analytics, maestro | | 25 | Org assessment, metadata review, permissions audit, Flow automation, Apex/LWC code review, release readiness, integration, marketing consent, Agentforce risk review, zero-trust maturity, DevSecOps pipeline, SOQL generation, Apex generation and test generation, operational T1/T2 runtime skills | 38 | Maestro routing (microsoft/m365/d365/Power Platform/Copilot), Entra Zero Trust & Conditional Access, Microsoft 365 Copilot readiness, Purview data security & compliance, Defender XDR SecOps, Intune endpoints, Teams collaboration, Exchange/SharePoint information governance, tenant governance, backup/BCDR & data resilience, licensing/EA optimization, Dataverse/DLP security, Power Platform ALM, Power Automate risk review, Copilot Studio governance, Fabric/Power BI governance, Fabric data engineering, Fabric analytics engineering, D365 Success by Design, SoD, data migration/cutover, finance close-to-report, supply chain, field service, customer service, sales ops, Customer Insights – Journeys, F&O developer/extensions, dual-write integration, Project Operations, Commerce, value realization, live-guard identity posture + Dataverse security role (read-only-runtime, Phase A), live-guard Dataverse record field update + sensitivity label apply (mutating-runtime, Phase B) | | 20 | Cloud-neutral lakehouse/AI board — maestro routing plus specialists for platform architecture, Unity Catalog governance, identity/network security, data protection & privacy, Lakeflow pipelines, streaming reliability, data quality & observability, SQL performance, AI/BI Genie, MLOps, GenAI agent engineering, GenAI evaluation & observability, developer platform (bundles/CI-CD), platform reliability, FinOps cost, and value realization — all static review. Plus the hand-authored -at-azure assets that predate the board: Unity Catalog governance, lakehouse engineering, and a Unity Catalog schema-scoped grant guard (mutating-runtime live-guard, Phase B) | | 28 | Cloud-neutral data-platform board — router plus review specialists across identity/access, governance & privacy, compliance evidence, Cortex AI security, BCDR resilience, and business value; plus snowflake-live-* mutating-runtime guards, each a single approved mutation behind a human gate. The Azure-scoped predecessors remain as deprecated assets |

🛡️ Live Guard skills — stop before you break prod

Live-guard skills enforce approval gates and rollback posture for irreversible operations:

🟧 AWS (5):

aws-live-deployment-guarded-operator — approval-gated generic live deployment actions with account/region confirmation
aws-live-iac-change-guard — CloudFormation/SAM/CDK/Terraform change set + drift + rollback posture enforcement
aws-live-pipeline-approval-operator — CodePipeline approval gating with exact stage and approver scope
aws-live-serverless-release-guard — Lambda alias/canary/linear rollout with alarm + rollback required
aws-live-ecs-rollout-guard — ECS/Fargate deployment circuit breaker, health check evidence, rollback path

🟩 GCP (6):

gcp-live-gke-rollout-guard — GKE deployment and node pool mutations, control-plane version gating
gcp-live-iam-policy-change-guard — IAM binding mutations, org policy changes, SA key creation — org-wide blast radius
gcp-live-kms-key-destruction-guard — Cloud KMS key version destruction — CMEK data permanently unrecoverable
gcp-live-cost-budget-action-guard — budget thresholds, CUD commitments, quota increases — financial authority gate
gcp-live-bigquery-dataset-deletion-guard — dataset deletion, table truncation, authorized view changes — irreversible data loss
gcp-live-cloud-run-traffic-migration-guard — Cloud Run revision traffic shifts, min-instances changes — production traffic blast radius

🟠 Alibaba Cloud (6):

alibaba-live-ack-rollout-guard — ACK deployment mutations, node pool scaling, cluster version upgrades
alibaba-live-ram-policy-change-guard — RAM policy/role mutations — account-wide blast radius, privilege escalation risk
alibaba-live-kms-key-mutation-guard — KMS key deletion/disable — encrypted data permanently inaccessible
alibaba-live-cost-budget-action-guard — budget threshold changes, Savings Plan purchases, RI commitments — financial authority gate
alibaba-live-oss-bucket-policy-guard — OSS bucket ACL/policy changes — public exposure or China data-residency violation
alibaba-live-rds-polardb-mutation-guard — RDS/PolarDB instance deletion, spec downgrade, backup policy removal — data loss risk

🔴 Huawei Cloud (6):

huawei-live-cce-rollout-guard — CCE deployment mutations, node pool upgrades, cluster version changes
huawei-live-iam-policy-change-guard — IAM policy/SCP mutations — account-wide blast radius, privilege escalation
huawei-live-kms-key-destruction-guard — DEW/KMS key deletion — CSMS secrets and DBSS-encrypted data permanently lost
huawei-live-cost-budget-action-guard — budget threshold changes, RI purchases, CUD commitments — financial authority gate
huawei-live-obs-bucket-policy-guard — OBS bucket ACL/policy changes — public exposure or data residency violation
huawei-live-gaussdb-mutation-guard — GaussDB/RDS instance deletion, spec downgrade, backup policy changes — data loss

🟦 Azure (7):

azure-live-aks-rollout-guard — PDB audit, rollout pause/undo, post-rollout health
azure-live-arm-deployment-stack-guard — what-if evidence, denySettings, PIM-gated delete
azure-live-app-service-slot-swap-guard — sticky-setting audit, traffic shifting, swap-back path
azure-live-keyvault-rotation-purge-guard — rotation policy, soft-delete/purge-protection, PIM gate
azure-live-pim-jit-activation-guard — eligible assignment audit, MFA gate, JIT revocation
azure-live-cost-budget-action-guard — budget mutation, GPU SKU policy, quota read-only
azure-live-entra-role-assignment-guard — permanent role assignment scope/principal audit, PIM-preference enforcement, Guest principal blocking

🟥 OCI (7):

oci-live-autonomous-db-lifecycle-guard — ADB scale/stop/clone/terminate with tag enforcement
oci-live-oke-rollout-guard — DevOps pipeline approval, PDB audit, rollout pause/undo
oci-live-resource-manager-stack-guard — plan-before-apply, drift detection, job-lock enforcement
oci-live-vault-key-destruction-guard — rotation vs. destruction separation, 7–30 day deletion window
oci-live-iam-policy-compartment-guard — MFA break-glass, dual-approval for tenancy-root changes
oci-live-cost-budget-runaway-guard — 3-tier budget management, GPU shape gate, ONS alert routing
oci-live-network-security-rule-guard — Security List/NSG rule capture, 0.0.0.0/0 detection, DB-subnet criticality, Path Analyzer gate

☸️ Kubernetes (5):

kubernetes-live-rbac-mutation-guard — escalate/bind/impersonate verb detection, wildcard blocking, pre-mutation state capture, rollback via YAML backup
kubernetes-live-admission-policy-guard — Kyverno/VAP mutation blast-radius, failureAction enforcement, PolicyException scope validation
kubernetes-live-mesh-policy-guard — Istio AuthorizationPolicy/PeerAuthentication traffic impact, PERMISSIVE→STRICT migration gating
kubernetes-live-network-policy-guard — CiliumNetworkPolicy/NetworkPolicy connectivity impact, metadata service egress blocking
kubernetes-live-argocd-sync-guard — AppProject blast-radius, sync impersonation identity review, sync-window change gating

Sample skills

🔐 skills/aws/aws-iam-least-privilege-review — Review AWS IAM policies and reduce unnecessary access.
🟦 skills/azure/azure-rbac-review — Review Azure RBAC assignments, scopes, and custom roles.
🟥 skills/oci/oci-autonomous-database-architect — Design and review Oracle Autonomous Database across OCI and multicloud options.
🟩 skills/gcp/gcp-gke-platform-operator — GKE Standard/Autopilot design with Day-0/Day-1 decisions, Workload Identity, and AI inference quickstart.
🟠 skills/alibaba/alibaba-china-compliance — MLPS 2.0, DSL, PIPL, and ICP filing compliance review for workloads in Alibaba Cloud China regions.
🔴 skills/huawei/huawei-secmaster-security-operations — Huawei SecMaster SIEM/SOAR, HSS, CFW, WAF posture hardening and incident triage.
💰 skills/finops/finops-cloud-price-advisor — Fetch live prices from AWS, Azure, and OCI public pricing APIs; estimate costs for live environments or prototypes.
📣 skills/marketing/marketing-pixel-data-leakage-review — Review advertising pixels and event tracking for PII/PHI leakage to ad networks.

Rule of thumb: if the asset teaches how to do a repeatable task, it is a skill.

🤖 Agents

735 agents matching the skill catalog — agents ship harness adapters and a hardened permission model.

| Provider | Count | Specialisations | | ------------------ | ----: | ----------------------------------------------------------------------------------- | | 🟩 GCP | 51 | advisory, live-guard operators, maestro router | | 🟧 AWS | 47 | advisory, execution, live-guard operators | | 🟠 Alibaba Cloud | 43 | advisory, live-guard operators, maestro router | | 🔴 Huawei Cloud | 43 | advisory, live-guard operators, maestro router | | 🟥 OCI | 39 | advisory, live-guard operators | | 🟦 Azure | 36 | advisory, live-guard operators | | ☸️ Kubernetes | 16 | RBAC review, workload identity, PSA, 5 live-guard operators, maestro router | | ☁️ OVHcloud | 6 | advisory, live KMS guard, maestro router | | 🌐 IONOS Cloud | 6 | advisory, live DB lifecycle guard, maestro router | | 🇫🇷 Scaleway | 6 | advisory, live Kapsule rollout guard, maestro router | | 🇩🇪 Hetzner Cloud | 6 | advisory, live firewall + server lifecycle guards, maestro router | | 💰 Contabo | 6 | advisory, live instance + storage guards, maestro router | | 🛡️ Kyverno | 1 | Admission policy review | | 🔄 Argo CD | 2 | GitOps review, live sync guard | | 🕸️ Istio | 7 | Maestro plus ambient, authorization, traffic resilience, Gateway API, upgrade, and dataplane review | | 🐝 Cilium | 1 | Network policy review | | 📡 OpenTelemetry | 1 | Collector config review | | 💡 Backstage | 1 | IDP scaffolder review | | 🔐 cert-manager | 1 | PKI certificate lifecycle review | | 🦅 Falco | 1 | runtime threat detection review | | 🔁 Flux CD | 1 | GitOps Kustomization/HelmRelease review | | 📊 Prometheus | 1 | alerting and cardinality review | | 🔏 Sigstore | 1 | supply-chain security review | | 🟩 Terraform | 9 | Engine-shared Terraform/OpenTofu board — one decision right each: plan blast radius, state reliability, estate reconciliation, supply-chain integrity, engine compatibility, policy evidence, execution governance, module contract, maestro | | 💸 FinOps | 4 | cross-cloud price advisor + experimental cost/economics agents | | 🐘 PHP | 5 | maestro + 4 static-review specialists: application security (session fixation, deserialization, file-upload), runtime EOL/OPcache/FPM hardening, Composer supply-chain audit, WordPress REST API/block-editor security — static-review only | | 🟣 .NET | 10 | C#/runtime, ASP.NET Core API & identity, EF Core data access, testing, NuGet supply chain, performance/AOT, OpenTelemetry, Aspire — static-review specialists + maestro router | | ☕ Java | 15 | JDK lifecycle & upgrade, concurrency/virtual threads, JVM performance & GC, container/K8s sizing, framework readiness (Spring Boot/Quarkus/Micronaut), Spring Security, deserialization/parser security, JPA/Hibernate performance, transaction & consistency, migration safety, Kafka reliability, resilience patterns, test architecture, app-server license exit — static-review specialists + maestro router | | 🟧 Kotlin | 16 | language/null-safety correctness, coroutines/Flow reliability, library API/ABI governance, Ktor + Kotlin-Spring backend readiness, kotlinx.serialization wire contracts, Java-to-Kotlin modernization, Android (architecture, Compose UI + accessibility, MASVS security/privacy, runtime performance), Kotlin Multiplatform (portfolio + interop), Gradle build engineering, dependency/release supply-chain integrity, test architecture — static-review specialists + maestro router | | 🐍 Python | 35 | 20 static-review specialists (application security, asyncio reliability, packaging/supply-chain, numerical correctness, typing contracts, web-service readiness, data-access/transactions, distributed-task reliability, testing quality, estate modernization, performance/memory, free-threading/no-GIL, native-extension interop, container/serverless, data-pipeline reliability, ML/AI production, observability, tooling/build, automation governance) routed by python-maestro-agent; plus a 15-agent live control plane (python-live-*, read-only-runtime + mutating-runtime) routed by python-live-governance-maestro-agent — mutating operators are live-guard gated (external signed approval bound to target, JIT credentials, pre-approved rollback, immutable audit event, fail-closed for R3+); ships definitions/contracts/evals, not a running control plane | | 🔷 TypeScript | 14 | maestro router + 13 static-review specialists: type soundness in shared/published code, runtime boundary contracts (parse-don't-validate at every external edge), module resolution and emit across consumer modes, Node execution compatibility (type stripping vs type checking), public API and declaration governance, build-graph performance (measurement-gated), static enforcement policy (compiler flags + typed lint), async contract reliability, package publication integrity, estate modernization sequencing, MCP tool contracts, business-critical automation governance, engineering economics — static-review only, no agent compiles, builds, publishes, or contacts a live system | | 🟤 NVIDIA | 12 | CUDA/GPU infrastructure, TensorRT/TensorRT-LLM, Triton serving, NeMo/NIM generative AI, agentic-AI platform, NGC supply chain, AI networking fabric, day-2 operations, GPU Operator on Kubernetes, model promotion gatekeeper — advisory + live-runtime gate + maestro router | | 📣 Marketing | 14 | 13 governance review agents + maestro router | | ⚖️ Legal | 13 | contract review, employment law risk, privacy & data protection, regulatory compliance, IP & open source, litigation & discovery hold, ethics & investigations, vendor/procurement risk, policy governance, public disclosure, counsel review, knowledge management | | 👥 HR | 15 | employee relations, workplace investigations, performance management, compensation & equity, benefits & payroll, recruiting & selection, workforce planning & RIF, leave & accommodation, learning policy, culture & DEI, people analytics, HRIS process controls, termination readiness, risk triage | | 🧪 QA | 10 | Playwright E2E review + execution, flakiness triage, coverage quality, CI test pipeline review, PLC control-logic safety, RPA workflow resilience — static-review + opt-in execution | | ☁️ Salesforce | 30 | 20 Wave 1 domain specialists (admin, dev, security, integration, Sales/Service/Marketing/Industry clouds, Agentforce, analytics, compliance) + 10 Wave 3 infrastructure security + DevSecOps agents — maestro router + live-guard authority gate | | 🔷 SAP | 40 | maestro + 39 specialists: S/4HANA transformation architecture, BTP account/entitlement governance, ABAP Cloud/RAP & CAP code review, Fiori/UI5 UX, finance FI-CO controls, supply chain IBP resilience, manufacturing execution risk, SuccessFactors HR process risk, MDG master data quality, custom-code remediation, Cloud ALM SRE, transport/release governance, read-only landscape + identity discovery; + 4 guarded-mutating-runtime live-guard operators (BTP entitlement, integration flow, role assignment, transport import) — advisory + live-guard-gated execution + maestro router | | Ⓜ️ Microsoft 365 / D365 | 40 | 5 maestro routers (microsoft, m365, d365, Power Platform, Copilot governance) + 31 static-review specialists + 2 read-only-runtime live-guards (Phase A: identity posture, Dataverse security role) + 2 mutating-runtime live-guards (Phase B: Dataverse record field update, sensitivity label apply) across M365 (Entra identity/Zero Trust, Copilot readiness, Purview data security & compliance, Defender XDR SecOps, Intune endpoints, Teams collaboration, Exchange/SharePoint information governance, tenant governance, backup/BCDR & data resilience, licensing/EA optimization), Power Platform (Dataverse security, ALM pipelines, automation risk, Copilot Studio governance), Fabric/Power BI (governance, data engineering, analytics engineering), D365 (Success by Design, SoD, data migration/cutover, finance, supply chain, field service, customer service, sales, Customer Insights – Journeys, F&O developer/extensions, dual-write integration, Project Operations, Commerce, value realization), and live-guard runtime (Phase A: identity posture, Dataverse security role; Phase B: Dataverse record field update guard, sensitivity label apply guard) | | 🧱 Databricks | 20 | Cloud-neutral board (static review): databricks-maestro-agent routes on seven axes (intent, business context, artifact type, blast radius, required evidence, implied runtime authority, specialist ownership) to platform architecture (account/workspace topology, metastore-per-region, serverless vs classic), Unity Catalog governance (three-level namespace, GRANT model, workspace-catalog binding, governed tags), identity/network security (SCIM, service principals, OAuth vs PAT, IP access lists, serverless egress, secret scopes), data protection & privacy (row filters, column masks, ABAC, classification, GDPR erasure via REORG/VACUUM, Delta Sharing egress, residency, CMK), Lakeflow pipeline engineering (Lakeflow Spark Declarative Pipelines, medallion, liquid clustering, Auto Loader), streaming reliability (checkpoints, state schema, watermarks, exactly-once, RocksDB), data quality & observability (expectations, Lakehouse Monitoring, drift, freshness), SQL performance (warehouse sizing, Photon, cache tiers, query profile), AI/BI Genie (Genie Agents, metric views, trusted assets), MLOps (MLflow 3, models in Unity Catalog, alias promotion, Model Serving, inference tables), GenAI agent engineering (Agent Framework, Databricks AI Search, MCP, Unity AI Gateway), GenAI evaluation & observability (MLflow Tracing, built-in judges, regression detection), developer platform (Declarative Automation Bundles, run-as, CLI/OAuth, Terraform, Git flows), platform reliability (system tables, job retries/timeouts, cluster policies, managed DR), FinOps cost (billing system tables, tag attribution, DBU uptime, budgets), and value realization (baseline-or-no-number, attribution limits, kill conditions). Azure-specific -at-azure agents that predate the board and remain alongside it: Unity Catalog governance and lakehouse engineering (ADLS Gen2 / Access Connector / VNet), both static review, plus 1 mutating-runtime live-guard (Phase B): a Unity Catalog schema-scoped grant guard (single GRANT to one principal, REVOKE rollback, written approval token + PREFLIGHT required) | | ❄️ Snowflake | 28 | Review board (static review): snowflake-maestro-agent routes to specialists for effective access computed transitively across grants and ownership, governance/privacy controls, independent compliance-evidence assurance, Cortex AI agent trust boundaries, BCDR (requested vs feasible vs proven RPO/RTO), and economic justification with veto authority. Live guards (mutating-runtime): the snowflake-live-* guards — RBAC grant, auth/network policy, warehouse cost, data-protection policy, pipeline/streaming change, and failover promotion — each one operation behind a written human gate with preflight, prior-state capture, verification, and rollback; never auto-dispatched. Azure-scoped predecessors are retained as deprecated | | 🔗 Cross-functional skills | 3 | legal-hr-routing-protocol, legal-hr-case-capsule, legal-hr-risk-taxonomy (protocol skills, not agents) |

⚖️ The Legal + HR cross-functional agentic ecosystem

Beyond cloud and platform agents, Vanguard Frontier ships a 28-agent cross-functional Legal + HR ecosystem plus 3 cross-functional protocol skills — proof that agentic coordination works across organizational boundaries, not just inside one cloud account.

Every Legal and HR agent is escalation-aware (knows when a matter must go to privileged counsel or a human owner), privacy-preserving (minimizes personal and sensitive data in every handoff), and audit-ready (emits the same structured verdict shape as the cloud live-guard agents). These agents advise on process and risk posture — they do not replace licensed legal counsel or qualified HR professionals, and they say so.

🐘 The PHP application review board

PHP is a widely used open-source, general-purpose scripting language built for the web — it powers a large share of the internet, including WordPress, and ships its own dependency manager, Composer. The board is a php-maestro router plus four static-review specialists covering application security (session fixation, insecure deserialization, and file-upload exploits), runtime end-of-life/OPcache/PHP-FPM configuration hardening, Composer supply-chain dependency audit, and WordPress REST API / block-editor security — every agent reads source, sanitized configuration, and dependency manifests only and never executes payloads, installs packages, or mutates a live runtime. These agents declare their own provider: php catalog value with a mix of ID prefixes (php-, composer-, wordpress-) reflecting the specific ecosystem tool each one reviews.

🟣 The .NET application review board

.NET is a free, cross-platform, open-source developer platform — runtime, libraries, and languages (C# is the most popular) — with ASP.NET Core as its lean, modular framework for modern cloud-based web services and EF Core as its lightweight, extensible data-access layer. The board is a dotnet-maestro router plus nine static-review specialists covering C#/runtime correctness, ASP.NET Core API architecture, identity and authorization, EF Core data access, test quality, CI/NuGet supply chain, performance/AOT/trimming, in-app OpenTelemetry wiring, and .NET Aspire cloud-native readiness — every agent reads source and sanitized configuration only and never builds, runs, migrates, or contacts a live system. These agents use provider: generic with a dotnet- ID prefix because .NET is a language/runtime, not a cloud provider — mirroring the existing non-cloud boards.

☕ The Java application review board

Java and the JVM run a large share of the enterprise back end — Spring Boot, Jakarta EE, Quarkus, and Micronaut services over Hibernate/JPA, Kafka, and relational databases, on JDKs whose support and licensing boundaries move. The board is a java-maestro router plus fourteen static-review specialists, each owning one distinct, evidence-gated decision: JDK-estate lifecycle and upgrade risk, virtual-thread adoption correctness, JVM performance and GC, containerized-JVM sizing, Spring/Quarkus/Micronaut production readiness, Spring Security authorization and Actuator exposure, untrusted-deserialization/parser RCE surface, JPA/Hibernate fetch performance, transaction and cross-resource consistency (dual-write → outbox), schema-migration deploy safety, Kafka delivery semantics, resilience-pattern composition, JVM test architecture, and the application-server license-exit portfolio decision. Every agent reads source and sanitized configuration only and never builds, runs, invokes a JDK, opens a database or broker connection, or contacts a live system. Two hardened refusal contracts anchor the board's honesty: the JDK agent never states a vendor lifecycle date from memory (it cites a verified, refreshable reference or returns unknown), and the performance and application-server-exit agents refuse to issue a GC recommendation or a payback number without supplied evidence — no hardcoded pricing or tenant data. These agents use provider: java with a java- ID prefix, mirroring the other non-cloud topical boards.

🟧 The Kotlin application review board

Kotlin runs across three estates that share a language but not a runtime — JVM back ends (Ktor and Kotlin-on-Spring services over coroutines and kotlinx.serialization), Android apps (lifecycle-aware architecture, Jetpack Compose, MASVS security), and Kotlin Multiplatform code shared to Apple and beyond. The board is a kotlin-maestro router plus fifteen static-review specialists, each owning one distinct decision: language and null-safety/interop correctness, coroutine and Flow reliability (structured concurrency, cancellation, dispatcher confinement, context propagation), public library API/ABI evolution, Ktor + Kotlin-Spring backend production readiness, kotlinx.serialization wire-contract safety, Java-to-Kotlin estate modernization sequencing, Android lifecycle-aware architecture, Jetpack Compose UI quality and accessibility, MASVS-aligned Android security and privacy, measured Android runtime performance, the Kotlin Multiplatform adopt-or-not portfolio decision, the KMP source-set/interop boundary (including Kotlin/Native), Gradle build engineering, dependency and release supply-chain integrity, and coroutine/Compose/KMP test architecture. Every agent reads source and sanitized configuration only and never builds, runs, deploys, signs, publishes, or contacts a live system; the maestro routes generic-JVM, cluster, observability, and signing concerns out to the Java, Kubernetes, OpenTelemetry/Prometheus, and Sigstore boards rather than answering them. These agents use provider: kotlin with a kotlin- ID prefix, mirroring the other non-cloud topical boards.

🐍 The Python application review + live governance board

Python spans application code, async runtimes, the packaging/supply-chain surface, data and ML pipelines, and the business-critical automation that quietly runs finance and operations. The board is the largest in the catalog and the only language/stack board with two planes, kept deliberately separate:

A 20-agent static-review board routed by python-maestro-agent: adversarial, evidence-first specialists, each owning one decision — application security (unsafe deserialization, dynamic execution, subprocess/shell injection, SSRF, path traversal, crypto misuse, each traced input→sink with a CWE), asyncio event-loop reliability, packaging and software-supply-chain integrity (dependency confusion, hash-checking, build isolation), numerical/scientific correctness (money-as-float, dtype coercion, reproducibility), gradual-typing contracts, web-service production readiness, database access and transaction scope, distributed-task idempotency, test-suite quality, runtime-estate modernization, performance/memory, free-threaded (no-GIL, PEP 703) adoption, native-extension interop, container/serverless runtime, data-pipeline reliability, ML/AI production, in-app observability, developer tooling/build, and business-critical automation governance. Every one reads source, sanitized configuration, and dependency manifests only — it never runs pip install, imports or executes the code, opens a connection, deploys, publishes, or edits files.
A 15-agent live control plane (python-live-*) routed by python-live-governance-maestro-agent, for controlled execution with provable accountability. Read-only-runtime agents (inventory, identity/authority, runtime observation, change planning, policy-gate evaluation, evidence sealing, continuous control testing, exception governance) observe and plan without mutating. Six mutating-runtime operators (code remediation, release control, data-change control, job control, model promotion, rollback/recovery) are live-guards: never auto-dispatched, each acting on exactly one bounded scope only behind an external signed approval bound to the exact target and plan digest, target-scoped just-in-time credentials, a pre-approved rollback, and an immutable audit event — failing closed when audit logging is unavailable for an R3+ action.

The board is a deliberate, governed exception to the "language/stack boards are static-review only" default, and it is scrupulous about the boundaries that matter: permission is not authority, execution is not approval, a green technical result is not business success, an emitted audit event is not proof an action was authorized, a control-to-framework mapping is not a compliance claim, and automation is not accountability. This repo ships the definitions, control/audit contracts, and eval fixtures — not a running control plane. The audit-log store, JIT issuance, approval system, and actual execution belong to the deploying organization's runtime, and regulatory/compliance classification stays its qualified owners' determination. These agents declare provider: python with a python- ID prefix; see docs/language-stack-boards.md, docs/evidence-output-spec.md, and docs/compliance/.

🔷 The TypeScript program and package board

TypeScript is not synonymous with React, and this board is deliberately not a second frontend board. Its unit of analysis is the TypeScript program and the published package: compiler configuration, module resolution and emit, the declaration surface, program-graph cost, Node execution, and the runtime boundaries that type erasure leaves undefended. The board is a typescript-maestro router plus thirteen static-review specialists, each owning one bounded decision: whether a type abstraction in shared or published code proves what its signature claims; where external data must be parsed rather than asserted; whether a package resolves and emits correctly for every consumer mode; whether the code runs on the target Node and is type-checked anywhere; whether a declaration change is breaking and what version it requires; what in the type graph costs measured build and editor time; what the static toolchain must prove and at what cost; whether every promise is awaited, cancellable, and bounded; who may publish and what ships in the tarball; how an estate migrates reversibly; whether a declared MCP tool contract matches its handler; whether a privileged script may run and under which controls; and what the supplied measurements make worth funding.

Two design constraints are load-bearing. First, compile-time is not runtime: TypeScript types are erased, so "it compiles" is evidence about the source and not about the payload — the runtime-boundary specialist exists precisely because a generated type is a claim about a producer rather than a check on data. Second, this repository contains no TypeScript program of its own, so no agent may ground a verdict in "the installed version"; every version-gated conclusion is preconditioned on evidence the user supplies, and an agent without it refuses rather than guesses. The boundary with the frontend board is drawn on artifact scope — application diffs and framework specifics stay there — and is written into both boards' refusal lists. Every agent reads source and sanitized configuration only and never compiles, builds, tests, publishes, deploys, or contacts a live system. These agents declare provider: typescript with a typescript- ID prefix; the full design record, including every candidate agent that was rejected and the reason, is in .claude/workflow/typescript-board/.

Ⓜ️ The Microsoft 365 / Dynamics 365 board

A maestro-routed board for the Microsoft business cloud (M365, Dynamics 365, Power Platform, and Copilot) — the SaaS surface, distinct from Azure infrastructure. A top-level microsoft-maestro routes to four sub-maestros (m365, d365, power-platform, copilot-governance) and refuses Azure-IaaS / cross-cloud work, deflecting it to the Azure/AWS/GCP maestros. Under them sit 31 static-review specialists + 2 read-only-runtime live-guards (Phase A) + 2 mutating-runtime live-guards (Phase B) organized into categories:

M365 — identity & Copilot: Entra Conditional Access / Zero Trust and Microsoft 365 Copilot readiness & data-exposure governance (the Copilot Zero Trust 7-layer / oversharing model).
Power Platform & Copilot Studio: environment + DLP + Dataverse security, ALM pipelines, automation (Power Automate) risk review, and Copilot Studio agent governance / ALM.
Analytics: Microsoft Fabric / Power BI semantic-model trust, RLS/OLS, and workspace governance.
Dynamics 365: Success by Design implementation governance, Finance & Operations segregation-of-duties, data migration & cutover, finance close-to-report, supply chain plan-to-produce, field service-to-cash, customer service & contact center, sales revenue operations, and license-to-value realization.

Static-review agents produce no tenant mutations; every verdict requires explicit human approval. The board also ships Phase A read-only-runtime live-guards (m365-live-identity-posture-guard-agent, d365-live-security-role-guard-agent) that read live tenant state without writing, and Phase B mutating-runtime live-guards (d365-live-record-field-update-guard-agent, m365-live-sensitivity-label-apply-guard-agent) that execute exactly one controlled WRITE operation (a named Dataverse field PATCH or a Graph sensitivity-label assignment to one driveItem) only after a written approval token, PREFLIGHT dry-run, and inverse-rollback path are confirmed — bulk edits, DELETE, ownerid/security-role fields, label-policy writes, and broad directory permissions are unconditionally denied. The board is grounded on Microsoft Learn and ships alongside 15 cross-functional protocol skills (provider: generic, recommendation-only) that orchestrate it across business processes — lead-to-cash, procure-to-pay, close-to-report, copilot-data-readiness, erp-crm-cutover, identity-to-data-access, and more.

🔷 The SAP S/4HANA + BTP board

SAP is a large enterprise ERP and business-cloud platform spanning S/4HANA (core ERP), the Business Technology Platform (BTP: integration, extension, and AI services), and a wide set of line-of-business modules. The board is an sap-maestro router plus 39 specialists spanning S/4HANA transformation architecture, BTP account/entitlement governance, ABAP Cloud/RAP and CAP code review, Fiori/UI5 UX, finance FI-CO controls, supply chain IBP resilience, manufacturing execution risk, SuccessFactors HR process risk, MDG master-data quality, custom-code remediation, Cloud ALM SRE/incident response, and transport/release governance — plus read-only landscape and identity/trust discovery agents. Four guarded-mutating-runtime live-guard operators (BTP entitlement, integration flow, role assignment, and transport import) execute exactly one approved change behind a written approval token, PREFLIGHT dry-run, and inverse-rollback path; every other agent is static-review and produces no system-of-record mutation.

Every agent ships:

📄 AGENT.md — harness-neutral contract with guarded response shape
🗂️ metadata.json — schema-validated catalog entry
🔌 Harness adapters — claude-code + codex (EU providers); all 7 adapters for established providers

agents/
├── accounting/       (14 agents — maestro + corporate accounting specialists: revenue recognition (ASC 606/IFRS 15), close cycle, consolidation/intercompany, fixed assets, lease accounting (ASC 842), hedge accounting, FX translation, tax provision, payroll, procure-to-pay, equity compensation, indirect tax/e-invoicing, business combinations — all advisory, no ledger writes)
├── alibaba/          (43 agents — advisory, live-guard operators, maestro)
├── argocd/           (2 agents — GitOps review, live sync guard)
├── aws/              (47 agents — advisory, execution, live-guard operators)
├── azure/            (36 agents — advisory, live-guard operators)
├── backstage/        (1 agent — IDP scaffolder review)
├── cert-manager/     (1 agent — PKI certificate lifecycle review)
├── cilium/           (1 agent — network policy review)
├── contabo/          (6 agents — advisory, live instance + storage guards, maestro)
├── cross-functional/ (1 agent — revenue-critical journey integrity review, spanning the checkout/billing path that no single provider board owns)
├── databricks/       (20 agents — cloud-neutral board: maestro + 16 static-review specialists across platform architecture, Unity Catalog governance, identity/network security, data protection & privacy, Lakeflow pipelines, streaming reliability, data quality, SQL performance, AI/BI Genie, MLOps, GenAI agent engineering, GenAI evaluation, developer platform, platform reliability, FinOps, value realization; + 2 Azure-specific static-review agents and 1 Unity Catalog schema-scoped grant guard — mutating-runtime live-guard, Phase B, -at-azure)
├── dotnet/           (10 agents — C#/runtime, ASP.NET Core, EF Core, testing, NuGet supply chain, performance/AOT, OpenTelemetry, Aspire — maestro + 9 specialists)
├── falco/            (1 agent — runtime threat detection review)
├── finance/          (8 agents — maestro + corporate finance specialists: variance