convez.ai
SkillAutomaçãoAvançado
lebobo88

ExecutiveSuite

lebobo88/ExecutiveSuite

C-suite multi-agente com orquestração hierárquica para decisões estratégicas corporativas

Ver no GitHub
InicialPluginScore · de 0 a 100

Sistema multi-agente C-suite completo para Claude Code com 20 agentes executivos especializados, 4 orquestradores (boardroom, M&A cockpit, crisis war-room, capital allocation) e 9 skills compartilhadas. Implementa protocolos corporativos estruturados conforme EU AI Act, NIST AI RMF e COSO ERM para suporte estratégico à tomada de decisão.

Instalação

$npx convez-ai lebobo88/ExecutiveSuite

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
2
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 3 meses
Versão
—
Issues abertas
11
Contribuidores
3

Casos de uso

  • Simulação de reuniões de conselho com consenso hierárquico entre perspectivas executivas
  • Análise estruturada de operações M&A com triangulação de riscos em 7 etapas
  • Gestão de crises com protocolo de war-room para cenários black swan
  • Alocação de capital com debate adversarial entre domains executivos

Veredito

ExecutiveSuite é um sistema robusto e governado para orquestração executiva em contextos corporativos complexos, com fundamentação em frameworks de regulação e risk management. Exige expertise em decisão estratégica para máximo aproveitamento — não é ferramenta para uso casual.

README

A tradução para o português deste README ainda está na fila.

ExecutiveSuite

Agents Skills Commands Platform License

A comprehensive multi-agent C-suite for Claude Code: 20 single-domain executive agents, 4 multi-agent orchestrators (synthetic boardroom, M&A cockpit, crisis war-room, capital allocation committee), 9 shared skills, and 9 slash commands. Grounded in a 37,000-word research document on corporate multi-agent AI systems for strategic decision support, information triage, and financial architecture integration.


Architecture

graph TB
    subgraph Commands ["Slash Commands (9)"]
        CMD1["/exec-brief"]
        CMD2["/board-meeting"]
        CMD3["/mna-review"]
        CMD4["/crisis-mode"]
        CMD5["/capital-decision"]
        CMD6["/quarterly-review"]
        CMD7["/decision-memo"]
        CMD8["/risk-stress"]
        CMD9["/executive-team"]
    end

    subgraph Orchestrators ["Multi-Agent Orchestrators (4)"]
        BR["boardroom\nHierarchical Consensus"]
        MNA["mna-cockpit\nM&A Triangulation (7-step)"]
        CW["crisis-warroom\nBlack Swan War-Room (6-step)"]
        CA["capital-allocation\nAdversarial Debate Protocol"]
    end

    subgraph Executives ["Single-Domain Executives (20)"]
        direction LR
        OPUS["CEO · CSO · CFO · CTO · CAIO\nCPO · CLO · Chief Risk Officer\n(opus tier — strategic)"]
        SONNET["COO · CRO · CIO · CDO · CISO\nCMO · CXO · CHRO · CSCO\nChief Comms · Chief Compliance\nChief Sustainability\n(sonnet tier — operational)"]
    end

    subgraph Skills ["Shared Skills (9)"]
        SK1["executive-protocol"]
        SK2["financial-frameworks"]
        SK3["ai-governance"]
        SK4["debate-protocol"]
        SK5["scenario-planning"]
        SK6["enterprise-risk"]
        SK7["mna-playbook"]
        SK8["crisis-response"]
        SK9["stakeholder-comms"]
    end

    CMD2 --> BR
    CMD3 --> MNA
    CMD4 --> CW
    CMD5 --> CA
    CMD1 --> Executives
    CMD6 --> Executives
    CMD7 --> Executives
    CMD8 --> Executives

    BR --> OPUS & SONNET
    MNA --> OPUS & SONNET
    CW --> OPUS & SONNET
    CA --> OPUS & SONNET

    OPUS --> Skills
    SONNET --> Skills

    style Commands fill:#f0f4ff,stroke:#0969da
    style Orchestrators fill:#fff8f0,stroke:#e36209
    style Executives fill:#f0fff4,stroke:#2ea44f
    style Skills fill:#fdf0ff,stroke:#7c3aed

Each orchestrator adopts executive perspectives in-process (sequential persona adoption, not subagent spawning), synthesizing cross-functional recommendations and surfacing tensions. This avoids the multi-agent failure modes (specification ambiguity, organizational breakdown, weak verification) cataloged in the research document.


Ecosystem

ExecutiveSuite is the "executive" squad in a mesh of nine sibling AI systems bound together by a tenth layer, AgentMesh. It runs standalone or as part of this larger multi-agent ecosystem. ExecutiveSuite enrolls into the mesh via the mesh-manifest.yaml at this repo's root.

graph LR
    subgraph Governance ["Governance & Substrate"]
        TE["<a href='https://github.com/lebobo88/TheEights'>TheEights</a>\nMemory · Audit · Identity\nGovernance · Self-Evolution\n(root of trust)"]
        AS["<a href='https://github.com/lebobo88/AgentSmith'>AgentSmith</a>\nInspection · N1..N10 invariants\nQuarantine · Sentinel"]
    end

    subgraph Orchestration ["Orchestration & Engineering"]
        HY["<a href='https://github.com/lebobo88/Hydra'>Hydra</a>\nLangGraph Multi-Squad Supervisor"]
        PP["<a href='https://github.com/lebobo88/pair-programmer'>pair-programmer</a>\nBest-of-N Engineering Harness\n(engineering squad)"]
    end

    subgraph Squads ["Squad Source-Packs"]
        ES["<b>ExecutiveSuite</b>\nC-Suite Decision Support\n(executive squad)"]
        RLM["<a href='https://github.com/lebobo88/RLM-Creative'>RLM-Creative</a>\nCreative / Media Studio\n(garland squad)"]
        MB["<a href='https://github.com/lebobo88/MarketBliss'>MarketBliss</a>\nEnterprise Marketing\n(five marketing-* squads)"]
        SEN["<a href='https://github.com/lebobo88/Senate'>Senate</a>\nLegal Curia — 12 Jurists\n(legal-compliance squad)"]
        XEN["<a href='https://github.com/lebobo88/Xenia-Support'>Xenia</a>\nSupport Hearth — 11 agents\n(customer-support squad)"]
    end

    subgraph Binding ["Binding Layer"]
        AM["<a href='https://github.com/lebobo88/AgentMesh'>AgentMesh</a>\nGoverned Control Plane\nregistry · lifecycle · observability\naudit · protocol edge · console"]
    end

    HY -- "dispatches goals" --> ES
    HY -- "dispatches goals" --> RLM
    HY -- "dispatches goals" --> MB
    HY -- "dispatches goals" --> SEN
    HY -- "dispatches goals" --> XEN
    HY -- "engineering" --> PP
    AS -- "inspects artifacts\nN1..N10 fail-closed" --> ES
    TE -- "evolves skills\nrubrics, agents" --> ES
    AM -- "enrolls + routes\n+ observes" --> ES
    AM -. "authority stays with" .-> TE
    TE --> AS --> HY

    style Governance fill:#fff0f0,stroke:#cf222e
    style Orchestration fill:#f0f4ff,stroke:#0969da
    style Squads fill:#f0fff4,stroke:#2ea44f
    style Binding fill:#fdf0ff,stroke:#7c3aed
SystemRoleRelationship to ExecutiveSuite
TheEightsShared memory / audit / identity / governance / self-evolution substrate; the root of trust.Proposes and manages evolution of ExecutiveSuite's skills, rubrics, and agent definitions through a governed propose/evaluate/commit cycle.
AgentSmithArtifact inspection, the N1..N10 fail-closed invariants, quarantine + sentinel (the Matrix warden).Validates ExecutiveSuite artifacts, enforces constitutional invariants, and can quarantine rogue artifacts.
HydraLangGraph multi-squad supervisor; routes, governs, synthesizes. Hosts the squads.Dispatches goals to ExecutiveSuite as the executive squad; delivers structured decision packets between squads.
pair-programmerBest-of-N engineering harness; Hydra's engineering squad.ExecutiveSuite's taxonomy, team pipelines, and debate protocols originated here.
RLM-CreativeCreative / media studio; the garland squad.Sibling squad; ExecutiveSuite's agent file format mirrors RLM-Creative.
MarketBlissEnterprise marketing platform; the five marketing-* squads.Sibling squad-pack under the same Hydra supervisor.
SenatePhD-level legal wing, "the Curia" — 12 jurists under the Twelve Tables, resolving by the Law of Citations, gatekept by the Tribune's Veto (HITL); the legal-compliance squad.Sibling squad; active legal/compliance counterpart to the C-suite's CLO/Chief Compliance roles.
XeniaCustomer-support "Hearth" — an 11-agent crew for ticket triage, recommendation, VoC, and approval-gated execution with WS-AUTH capability enforcement; the customer-support squad.Sibling squad; active customer-support counterpart.
AgentMeshThe thin, governed control plane binding all nine systems.ExecutiveSuite enrolls via mesh-manifest.yaml; AgentMesh routes and observes but enforces no governance of its own.

AgentMesh — binding control plane

AgentMesh is the tenth layer: the thin, governed control plane that unifies the nine sibling systems behind ONE registry (SQLite ~/.agentmesh/state.db; sole writer of ~/.hydra/backends.json), ONE lifecycle supervisor (Win32 Job Objects + crash-loop breaker + health probes), ONE observability plane (OTEL + structured logs), ONE federated read-only audit timeline (stitched from TheEights/AgentSmith/Hydra chains), ONE external protocol edge (A2A v0.3/v1.0, REST, MCP-over-HTTP), and ONE operator web console.

ExecutiveSuite enrolls into the mesh by shipping the mesh-manifest.yaml at this repo's root. Enrollment is fail-closed: JSON-Schema validation against AgentMesh/mesh-manifest.schema.json, constitution attestation (via TheEights), and AgentSmith structural inspection must all pass. AgentMesh's manifest declares the executive_suite backend, the es.ping health probe, the nine es.* MCP tools, and the governance routing (constitution attestation and audit federation flow through TheEights, not AgentMesh).

AgentMesh enforces no governance of its own — authority stays with TheEights → AgentSmith → Hydra (precedence order). It routes and observes; it does not arbitrate.


Quick Start

Once installed (see Installation), from any Claude Code session:

/executive-team                                # See the full roster
/exec-brief cfo Evaluate the Ohio plant expansion
/board-meeting Should we acquire CompetitorX?
/capital-decision $50M brand campaign vs. R&D investment
/mna-review Bolt-on of regional supplier
/crisis-mode red Ransomware on production ERP
/quarterly-review Q3 2026
/decision-memo Should we delay launch by 6 weeks?
/risk-stress

Roster

Executives (20)

AgentTitleModelKey Skills
ceoChief Executive Officeropusexecutive-protocol
csoChief Strategy Officeropusexecutive-protocol, scenario-planning
cfoChief Financial Officeropusexecutive-protocol, financial-frameworks
ctoChief Technology Officeropusexecutive-protocol
caioChief AI Officeropusexecutive-protocol, ai-governance
cpoChief Product Officeropusexecutive-protocol
cloChief Legal Officer / General Counselopusexecutive-protocol
chief-risk-officerChief Risk Officer (Enterprise Risk)opusexecutive-protocol, enterprise-risk, scenario-planning
cooChief Operating Officersonnetexecutive-protocol
croChief Revenue Officersonnetexecutive-protocol
cioChief Information Officersonnetexecutive-protocol
cdoChief Data Officersonnetexecutive-protocol
cisoChief Information Security Officersonnetexecutive-protocol
cmoChief Marketing Officersonnetexecutive-protocol
cxoChief Customer Experience Officersonnetexecutive-protocol
chief-communications-officerChief Communications Officersonnetexecutive-protocol, stakeholder-comms
chroChief Human Resources Officersonnetexecutive-protocol
chief-compliance-officerChief Compliance Officersonnetexecutive-protocol, ai-governance
cscoChief Supply Chain Officersonnetexecutive-protocol
chief-sustainability-officerChief Sustainability Officersonnetexecutive-protocol

Orchestrators (4)

AgentPurposeModelMax TurnsKey Skills
boardroomSynthetic boardroom facilitator — hierarchical consensusopus40executive-protocol, debate-protocol
mna-cockpitM&A opportunity triangulation — 7-step workflowopus50executive-protocol, mna-playbook, financial-frameworks, debate-protocol
crisis-warroomBlack swan capital preservation — 6-step workflowopus50executive-protocol, crisis-response, financial-frameworks, scenario-planning
capital-allocationCapital allocation committee — CFO-led adversarial debateopus30executive-protocol, financial-frameworks, debate-protocol

When to use which orchestrator

flowchart TD
    Q{"What kind of\ndecision?"}
    Q -->|"Cross-functional\nstrategy"| BR["<b>boardroom</b>\n/board-meeting"]
    Q -->|"Acquisition or\ndivestiture"| MNA["<b>mna-cockpit</b>\n/mna-review"]
    Q -->|"Crisis or\nblack swan"| CW["<b>crisis-warroom</b>\n/crisis-mode"]
    Q -->|"Capital allocation\ntrade-off"| CA["<b>capital-allocation</b>\n/capital-decision"]
    Q -->|"Single-domain\nquestion"| EB["<b>Any executive</b>\n/exec-brief"]
    Q -->|"Periodic\nreview"| QR["<b>Full C-suite</b>\n/quarterly-review"]

    style Q fill:#f0f4ff,stroke:#0969da
    style BR fill:#fff8f0,stroke:#e36209
    style MNA fill:#fff8f0,stroke:#e36209
    style CW fill:#fff0f0,stroke:#cf222e
    style CA fill:#fff8f0,stroke:#e36209
    style EB fill:#f0fff4,stroke:#2ea44f
    style QR fill:#f0fff4,stroke:#2ea44f

Skills (9)

SkillPurposeUsed By
executive-protocolMemo templates, board meeting format, RACI, OKR, escalation rules, dissent formatAll agents
financial-frameworksWACC, NPV, IRR, payback, EVA, real-options, Monte Carlo, covenant/liquidity checksCFO, mna-cockpit, capital-allocation, crisis-warroom
ai-governanceEU AI Act Article 9, NIST AI RMF, ISO/IEC 42001, model cards, HITL policyCAIO, chief-compliance-officer
debate-protocolAdversarial red-team debate: specification, briefs, cross-examination, adjudicationboardroom, mna-cockpit, crisis-warroom, capital-allocation
scenario-planning2x2 matrices, Monte Carlo setup, sensitivity/tornado, decision trees, war-games, reverse stressCSO, chief-risk-officer, crisis-warroom
enterprise-riskCOSO ERM 2017 + ISO 31000, risk taxonomy, appetite, KRIs, Bow-Tie, three-lines-of-defensechief-risk-officer
mna-playbookDeal thesis, screening, financial/commercial/legal diligence, valuation, integration playbookmna-cockpit
crisis-responseCrisis classification, triggers, escalation tiers, rapid liquidity stress, holding statements, AARcrisis-warroom
stakeholder-commsMendelow map, audience-specific frameworks, board decks, investor narrative, crisis statementschief-communications-officer

Slash Commands (9)

CommandUsageRoutes To
/exec-brief/exec-brief <slug> <question>Single executive by slug
/board-meeting/board-meeting [--format brief|strategic] <topic>boardroom orchestrator
/mna-review/mna-review <deal description>mna-cockpit orchestrator
/crisis-mode/crisis-mode <yellow|orange|red> <event>crisis-warroom orchestrator
/capital-decision/capital-decision <decision frame>capital-allocation orchestrator
/quarterly-review/quarterly-review Q[N] [YYYY]All 20 executives sequentially
/decision-memo/decision-memo [--exec slug|--board] <question>Single exec or boardroom
/risk-stress/risk-stress [--scenario base|mild|severe]CFO + Chief Risk + Chief Sustainability
/executive-team/executive-teamInformational roster listing

How It Works

Subagent Architecture

Each agent is a standard Claude Code subagent — a Markdown file with YAML frontmatter (name, description, model, maxTurns, skills). Agents are invoked by name via the Agent tool or routed automatically when their description matches the user's task.

Skills are loaded into context when listed in an agent's skills: frontmatter. The executive-protocol skill provides the memo, board-meeting, RACI, OKR, and escalation templates that every agent uses.

In-Process Persona Orchestration

The orchestrators (boardroom, mna-cockpit, crisis-warroom, capital-allocation) impersonate multiple executive perspectives in-process — sequential persona adoption per role, with synthesized output. They do not spawn subagents. This mitigates the multi-agent failure modes (specification ambiguity, organizational breakdown, weak verification) cataloged in the research document.

sequenceDiagram
    participant U as User
    participant CMD as /board-meeting
    participant BR as boardroom agent
    participant CEO as CEO perspective
    participant CFO as CFO perspective
    participant CTO as CTO perspective
    participant CRO as Chief Risk perspective

    U->>CMD: /board-meeting "Should we enter the EU market?"
    CMD->>BR: Route to boardroom agent
    BR->>CEO: Adopt CEO perspective
    CEO-->>BR: Strategic alignment assessment
    BR->>CFO: Adopt CFO perspective
    CFO-->>BR: Financial analysis (NPV, IRR, WACC)
    BR->>CTO: Adopt CTO perspective
    CTO-->>BR: Technical feasibility & platform readiness
    BR->>CRO: Adopt Chief Risk perspective
    CRO-->>BR: Risk register & stress scenarios
    BR->>BR: Synthesize, surface tensions, flag HITL gates
    BR-->>U: Board minutes + recommendation + dissent record

Financial Hardcoding Directive

Per the research document, agents treat the following as first-class deterministic tools, not informal guidance:

  • WACC computation (capital structure + market inputs)
  • NPV, IRR, payback, economic profit (investment analysis)
  • Real-options valuation (binomial lattice or Monte Carlo simulation)
  • Monte Carlo scenario engines (probability-weighted outcomes)
  • Covenant / leverage / liquidity checkers (hard guardrails)

Hard guardrails are enforced before any agent recommends action:

GuardrailThreshold
Minimum IRRWACC + risk premium (0% low, 2% med, 5% high, 10%+ venture/M&A)
Maximum leverageNet debt / EBITDA ≤ 3.0x (board-set)
Covenant headroom≥ 20%
Liquidity runway≥ 12 months (base), ≥ 6 months (stress)
Counterparty concentrationNo single counterparty > 10% of receivables
Sanctions exposureZero tolerance

Governance Posture

  • Financial Hardcoding Directive — WACC / NPV / IRR / real-options / Monte Carlo / covenant / liquidity checks as first-class tools (see financial-frameworks skill)
  • EU AI Act Article 9 — risk classification + lifecycle gates + audit trail (see ai-governance skill)
  • NIST AI RMF — GOVERN / MAP / MEASURE / MANAGE function map
  • COSO ERM 2017 + ISO 31000 — enterprise risk framework
  • HITL gates — every high-impact recommendation surfaces required human approvals
  • Audit trail — every recommendation traceable to source data, exec persona, framework applied
  • Dissent preservation — dissenting opinions recorded verbatim, never paraphrased away

Output Structure

Executive artifacts are saved to output/<domain>/<topic>-YYYY-MM-DD.md:

output/
  strategy/          # ceo, cso
  finance/           # cfo, capital-allocation
  revenue/           # cro
  risk/              # chief-risk-officer
  operations/        # coo
  supply-chain/      # csco
  technology/        # cto
  it/                # cio
  data/              # cdo
  ai/                # caio
  security/          # ciso
  product/           # cpo
  marketing/         # cmo
  customer/          # cxo
  communications/    # chief-communications-officer
  people/            # chro
  legal/             # clo
  compliance/        # chief-compliance-officer
  sustainability/    # chief-sustainability-officer
  board/             # boardroom
  mna/               # mna-cockpit
  crisis/            # crisis-warroom

The output/ directory is relative to the project Claude Code was launched from. Each project accumulates its own artifact tree.


Installation

Project scope (default)

Clone the repository and run Claude Code from its directory — the .claude/ folder is auto-discovered:

git clone https://github.com/lebobo88/ExecutiveSuite.git
cd ExecutiveSuite

User scope (available in every project)

Install agents, skills, and commands into ~/.claude/ so they are available globally.

Windows — symlink (recommended for live updates)

Requires either an elevated PowerShell or Windows Developer Mode enabled (Settings > System > For developers > Developer Mode). Run from the cloned ExecutiveSuite directory:

$src = Join-Path (Get-Location) ".claude"
$dst = Join-Path $env:USERPROFILE ".claude"

foreach ($d in 'agents','skills','commands') {
  Get-ChildItem "$src\$d" | ForEach-Object {
    $linkPath = Join-Path "$dst\$d" $_.Name
    if (Test-Path $linkPath) { Remove-Item $linkPath -Recurse -Force }
    New-Item -ItemType SymbolicLink -Path $linkPath -Target $_.FullName | Out-Null
  }
}

Verify:

$dst = Join-Path $env:USERPROFILE ".claude"
foreach ($d in 'agents','skills','commands') {
  $n = (Get-ChildItem "$dst\$d" |
        Where-Object { $_.LinkType -eq 'SymbolicLink' -and $_.Target -like '*ExecutiveSuite*' }).Count
  "$d : $n symlinks"
}
# Expected: agents : 24, skills : 9, commands : 9

Windows — copy (no Developer Mode / Admin needed)

Edits in the repo require re-running this to propagate:

$src = Join-Path (Get-Location) ".claude"
$dst = Join-Path $env:USERPROFILE ".claude"

Copy-Item "$src\agents\*"   "$dst\agents\"   -Recurse -Force
Copy-Item "$src\skills\*"   "$dst\skills\"   -Recurse -Force
Copy-Item "$src\commands\*" "$dst\commands\" -Recurse -Force

macOS / Linux

src="$(pwd)/.claude"   # run from the cloned ExecutiveSuite directory
dst="$HOME/.claude"
for d in agents skills commands; do
  for f in "$src/$d"/*; do
    ln -snf "$f" "$dst/$d/$(basename "$f")"
  done
done

What is NOT installed globally

  • settings.json — project-scoped (output root + status line). Do not promote to user scope.
  • CLAUDE.md — project contract. Promoting it would inject ExecutiveSuite framing into unrelated projects.

Rollback (user scope)

Removes only ExecutiveSuite assets from ~/.claude/ (links or copies); leaves the source repo untouched:

$dst = Join-Path $env:USERPROFILE ".claude"

# Symlinks: filter by target
foreach ($d in 'agents','skills','commands') {
  Get-ChildItem "$dst\$d" |
    Where-Object { $_.LinkType -eq 'SymbolicLink' -and $_.Target -like '*ExecutiveSuite*' } |
    ForEach-Object { Remove-Item $_.FullName -Force }
}

# Copies: remove by name
$execAgents = 'ceo','cso','coo','cfo','cro','chief-risk-officer','cto','cio','cdo','caio','ciso','cpo','cmo','cxo','chief-communications-officer','chro','clo','chief-compliance-officer','csco','chief-sustainability-officer','boardroom','mna-cockpit','crisis-warroom','capital-allocation'
$execAgents | ForEach-Object { Remove-Item "$dst\agents\$_.md" -ErrorAction SilentlyContinue }

$execSkills = 'executive-protocol','financial-frameworks','ai-governance','debate-protocol','scenario-planning','enterprise-risk','mna-playbook','crisis-response','stakeholder-comms'
$execSkills | ForEach-Object { Remove-Item "$dst\skills\$_" -Recurse -Force -ErrorAction SilentlyContinue }

$execCmds = 'exec-brief','board-meeting','mna-review','crisis-mode','capital-decision','quarterly-review','decision-memo','risk-stress','executive-team'
$execCmds | ForEach-Object { Remove-Item "$dst\commands\$_.md" -ErrorAction SilentlyContinue }

Provenance

  • Research basis: Corporate Multi-Agent AI Systems for C-Suite Strategic Decision Support, Information Triage, and Financial Architecture Integration.md (included in this repository)
  • Agent format: Mirrors RLM-Creative (media-vertical C-suite agents, generalized here to enterprise)
  • Orchestration patterns: Informed by pair-programmer (taxonomy, teams, debate harness)

Related Projects

ProjectRoleLink
TheEightsShared memory / audit / identity / governance / self-evolution substrate — the root of trustgithub.com/lebobo88/TheEights
AgentSmithArtifact inspection, N1..N10 fail-closed invariants, quarantine + sentinel (the Matrix warden)github.com/lebobo88/AgentSmith
HydraLangGraph multi-squad supervisor — dispatches goals to ExecutiveSuitegithub.com/lebobo88/Hydra
pair-programmerBest-of-N engineering harness — taxonomy, teams, best-of-N judging (engineering squad)github.com/lebobo88/pair-programmer
RLM-CreativeCreative / media studio — domain-specific C-suite (garland squad)github.com/lebobo88/RLM-Creative
MarketBlissEnterprise marketing platform — the five marketing-* squadsgithub.com/lebobo88/MarketBliss
SenatePhD-level legal Curia — 12 jurists under the Twelve Tables (legal-compliance squad)github.com/lebobo88/Senate
XeniaCustomer-support Hearth — 11-agent crew with WS-AUTH enforcement (customer-support squad)github.com/lebobo88/Xenia-Support
AgentMeshThe governed control plane binding all nine systems — registry, lifecycle, observability, audit, protocol edgegithub.com/lebobo88/AgentMesh

Extending

  • Add an industry vertical: Clone this repo, override agent files in .claude/agents/ for your domain. Project scope wins per Claude Code subagent precedence.
  • Add a new executive: Drop a <slug>.md into .claude/agents/ following the existing template format (YAML frontmatter + persona + responsibilities + decision framework + constraints).
  • Add a workflow: Drop a <name>.md into .claude/commands/ with a description: frontmatter and a body describing the steps.

License

MIT

Ver o original no GitHub

Bom para

Relacionados