boardroom-agent-state

name	boardroom-agent-state
description	Enrich boardroom agent JSON-LD state files with legend-derived domain knowledge, skills, persona, and language from the boardroom-agents spec
license	MIT
metadata	{"author":"ASISaga","version":"2.0","category":"boardroom","role":"agent-state-specialist"}
allowed-tools	Bash(python:*) Read Edit

Boardroom Agent State Skill

Role: Boardroom Agent State Enrichment Specialist
Scope: boardroom/mind/*/Manas/ — agent Manas (memory) and boardroom/mind/*/Buddhi/ — agent Buddhi (intellect)
Version: 2.0

Purpose

Orchestrate context enrichment across all boardroom agents. Each agent has a dedicated skill that contains the precise legend-derived domain_knowledge, skills, persona, and language values for that agent. Use this skill as the entry point to discover the right agent skill, or for cross-agent operations such as full-roster validation.

Agent Roster

Role	Legend	Manas file	Dedicated skill
CEO	Steve Jobs	boardroom/mind/ceo/Manas/ceo.jsonld	.github/skills/boardroom-agent-state-ceo/SKILL.md
CFO	Warren Buffett	boardroom/mind/cfo/Manas/cfo.jsonld	.github/skills/boardroom-agent-state-cfo/SKILL.md
COO	W. Edwards Deming	boardroom/mind/coo/Manas/coo.jsonld	.github/skills/boardroom-agent-state-coo/SKILL.md
CMO	Seth Godin	boardroom/mind/cmo/Manas/cmo.jsonld	.github/skills/boardroom-agent-state-cmo/SKILL.md
CHRO	Peter Drucker	boardroom/mind/chro/Manas/chro.jsonld	.github/skills/boardroom-agent-state-chro/SKILL.md
CTO	Alan Turing	boardroom/mind/cto/Manas/cto.jsonld	.github/skills/boardroom-agent-state-cto/SKILL.md
CSO	Sun Tzu	boardroom/mind/cso/Manas/cso.jsonl ¹	.github/skills/boardroom-agent-state-cso/SKILL.md
Founder	Paul Graham	boardroom/mind/founder/Manas/founder.jsonld	.github/skills/boardroom-agent-state-founder/SKILL.md

When to Use This Skill

• Single agent enrichment → use the agent's dedicated skill listed in the roster above
• Full-roster validation → use the validation workflow in this skill
• Adding a new agent → create a new dedicated skill, then add the agent to the roster table

¹ The CSO agent uses .jsonl (newline-delimited JSON) rather than .jsonld.

Context Layer Schema

The context object is immutable (read-only by BoardroomStateManager). Required fields:

Field	Constraint
name	Legend's full name (string)
fixed_mandate	Domain mandate (matches CXO_DOMAINS[id]["domain"])
core_logic	Defining maxim — 1–3 sentences from the legend
immutable_constraints	List of 3 read-only guiding principles

Enrichment fields (optional per schema validator, required per this spec):

Field	Constraint
domain_knowledge	List of 4–5 deep expertise areas
skills	List of 4–5 concrete capabilities
persona	2–4 sentence identity description
language	2–3 sentence vocabulary/style description

Content Layer Company/Product State Schema

The content.company_state and content.product_state objects are validated by BoardroomStateManager against _AGENT_PERSPECTIVE_KEYS and require:

entity_id, entity_name, perspective, domain_knowledge, skills, persona, language, software_interfaces, current_signals

→ See src/business_infinity/boardroom.py → BoardroomStateManager._AGENT_PERSPECTIVE_KEYS

Full-Roster Validation

# Validate all agent states and Buddhi files load correctly
PYTHONPATH=/tmp/aos_mock:src python3 -m pytest tests/ -q -k "boardroom"

# Spot-check context enrichment and Buddhi completeness across all agents
PYTHONPATH=/tmp/aos_mock:src python3 - <<'PY'
from business_infinity.boardroom import BoardroomStateManager
for agent_id in BoardroomStateManager.get_registered_agent_ids():
    ctx = BoardroomStateManager.load_agent_context(agent_id)
    for field in ("domain_knowledge", "skills", "persona", "language"):
        assert field in ctx, f"{agent_id} context missing '{field}'"
    buddhi = BoardroomStateManager.load_agent_buddhi(agent_id)
    assert buddhi["agent_id"] == agent_id
    print(f"✓ {agent_id}: {ctx['name']} — Manas and Buddhi OK")
PY

Tool Integration

# Run all tests
PYTHONPATH=/tmp/aos_mock:src python3 -m pytest tests/ -q

# Lint boardroom module
PYTHONPATH=/tmp/aos_mock:src python3 -m pylint src/business_infinity/boardroom.py --disable=C0114,C0115,C0116,E0401

Related Documentation

→ Spec: .github/specs/boardroom-agents.md — Legend archetypes, domain knowledge, and JSON-LD schema
→ State manager: src/business_infinity/boardroom.py → BoardroomStateManager
→ Manas files: boardroom/mind/*/Manas/
→ Buddhi files: boardroom/mind/*/Buddhi/buddhi.jsonld
→ Shared state: boardroom/state/ — collective boardroom state
→ MVP spec: .github/specs/mvp.md — C-suite agent roster
→ Repository spec: .github/specs/repository.md

---

Version: 3.0 — Updated to mind/Manas/Buddhi architecture
Last Updated: 2026-04-03