Menú
Cross-harness agent patterns — standardize agent execution across different coding assistants
Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.
Basado en la clasificación ocupacional SOC
| name | ecc-harness-patterns |
| description | Cross-harness agent patterns — standardize agent execution across different coding assistants |
Inspired by: ECC — the definitive cross-harness agent orchestration system
Purpose: Integrate multi-harness patterns from ECC into local-llm-server's agent orchestration layer.
ECC is a production-grade system supporting 7+ agent harnesses (Claude Code, Cursor, Codex, OpenCode, Gemini, Zed, GitHub Copilot) with:
# agents/harness_adapter.py
class HarnessAdapter:
HARNESSES = {
"claude_code": {"context_key": "workspace", "supports": ["streaming"]},
"cursor": {"context_key": "editor", "supports": ["streaming", "streaming_chunks"]},
"codex": {"context_key": "project", "supports": ["completion"]},
}
def normalize_request(self, harness_id: str, request: dict) -> dict:
"""Convert harness-native request to local-llm-server format"""
ECC's /hooks/session-* pattern applied to local-llm-server:
.claude/hooks/
├── session-start/
│ ├── 10-detect-harness.sh # Identify active harness
│ ├── 20-load-harness-prefs.sh # Load harness-specific settings
│ └── 30-emit-telemetry.sh # Send harness telemetry
├── session-pause/
│ └── 10-checkpoint-state.sh
└── session-stop/
└── 10-summarize-session.sh
Extend router/model_router.py to consider harness capabilities:
class CrossHarnessRouter(ModelRouter):
def select_model(self, task: TaskRequest, harness: str) -> str:
"""Route considering harness limits and preferences"""
# Claude Code + reasoning task → deepseek-r1 (powerful)
# Cursor + quick fix → qwen3-coder (fast)
# Codex + completion → fastest available
Track which harnesses are active and their performance:
# .claude/state/harness-registry.json
{
"active_harnesses": ["claude_code", "cursor"],
"session_history": [
{
"harness": "claude_code",
"session_id": "sess_123",
"model": "deepseek-r1:32b",
"duration_sec": 1847,
"tasks_completed": 12,
"success_rate": 0.92
}
]
}
agents/harness_adapter.py — adapter pattern for harness differences.claude/hooks/ — add session lifecycle managementrouter/model_router.py — add cross-harness routing logicservices/harness_registry.py — track active harnesses and performancebackend/server.py — emit harness metadata in responsestests/test_harness_adapter.py, tests/test_cross_harness_router.py.claude/skills/ecc-harness-patterns/SKILL.md (this file)agents/harness_adapter.py (new)services/harness_registry.py (new).claude/hooks/session-start/10-detect-harness.sh (new)router/model_router.py (extend)docs/architecture/cross-harness-support.md (new)docs/architecture/agent-orchestration.mddocs/adrs/002-model-routing.mdOnboard a new client/company onto this platform's real Agentic OS: create the company record, scan its websites/repos, auto-provision specialist agents, activate its 24x7 agency runtime, and know exactly which "OS" building blocks (memory, integrations, dashboard) already exist versus which are roadmap gaps. ADAPTED FROM: a third-party giveaway skill ("agentic-os-installer" by Gennaro Santoro / Operations Heroes) that described a generic vault + Google-suite + skill-pack installer. That skill's product (Obsidian vault, Gmail/Calendar/ Drive wiring, "skill packs") does not exist in this repo and its promotional content (Skool community link) does not belong here. This is a clean-room rewrite that keeps the useful idea — "stand up a working agency OS for a client from a short checklist" — and maps every step to the real module that already implements it in this codebase, per CLAUDE.md architecture rules.
Agile sprint planning, velocity tracking, and burndown metrics for agent-managed projects
Initiative-level portfolio management with dependency tracking and milestone coordination
AI-assisted engineering impact analysis — productivity metrics and code quality insights
Temporal context graph for agent memory — track entity relationships and state changes over time
Andrej Karpathy's coding guidelines for AI agents — concise, correct, and well-tested code