agent-orchestration-governance
Handles governance and orchestration for hushh-research multi-agent systems.
Install
mkdir -p .claude/skills/agent-orchestration-governance && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/13011" && unzip -o skill.zip -d .claude/skills/agent-orchestration-governance && rm skill.zipInstalls to .claude/skills/agent-orchestration-governance
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Use when changing repo-scoped Codex custom agents, subagent concurrency or depth, delegation policy, or handoff verification rules in hushh-research.Key capabilities
- →Add or tighten repo-scoped custom agents
- →Change subagent concurrency or depth limits
- →Define authority boundaries and child handoff shape
- →Keep agents thin and routed through existing skills
- →Preserve delegation checkpoint and truth-first handoff shape
How it works
The skill modifies configuration files and documentation to govern agent orchestration, ensuring agents adhere to defined limits, delegation policies, and handoff rules.
Inputs & outputs
When to use agent-orchestration-governance
- →Updating subagent concurrency limits
- →Changing delegation policy
- →Defining agent handoff rules
- →Governance for repo-scoped agents
About this skill
Hussh Agent Orchestration Governance Skill
Purpose and Trigger
- Primary scope:
agent-orchestration-governance-intake - Trigger on repo-scoped custom agents,
.codex/config.tomllimits, delegation policy, child handoff contracts, or workflow orchestration changes. - Avoid overlap with
codex-skill-authoringfor generic skill taxonomy andrepo-contextfor broad repository intake.
Coverage and Ownership
- Role:
owner - Owner family:
agent-orchestration-governance
Owned repo surfaces:
.codex/agents.claude/agents.codex/config.toml.codex/skills/agent-orchestration-governance.codex/workflows/agent-orchestration-governanceAGENTS.mddocs/reference/operations/README.mddocs/reference/operations/coding-agent-mcp.md
Non-owned surfaces:
repo-contextcodex-skill-authoringrepo-operationsfuture-planner
Do Use
- Adding or tightening repo-scoped custom agents.
- Changing subagent concurrency/depth limits or delegation policy.
- Defining authority boundaries, parent-only actions, or child handoff shape.
- Keeping agents thin and routed through existing owner/spoke skills.
Do Not Use
- Broad repo scans that should start with
repo-context. - Generic skill creation or taxonomy work that belongs to
codex-skill-authoring. - Domain implementation after the correct owner lane is clear.
- Recursive multi-agent expansion beyond bounded defaults without evidence.
Read First
AGENTS.md.codex/skills/agent-orchestration-governance/references/delegation-contract.md.codex/skills/codex-skill-authoring/references/skill-contract.md.codex/skills/codex-skill-authoring/references/truth-first-operating-kernel.mddocs/reference/operations/coding-agent-mcp.md.codex/skills/agent-orchestration-governance/references/anti-rationalization.md
Workflow
- Verify that a custom agent is justified; prefer skills/workflows when role specialization is not needed.
- Keep the fleet at the curated sweet spot: broad read-only evidence lanes, not one agent per skill.
- Preserve the repo-wide delegation checkpoint and truth-first handoff shape in
AGENTS.mdanddelegation-contract.md. - Keep custom-agent TOML files thin: role, sandbox, nicknames, root craft/bacterial-architecture inheritance hook, concise instructions, and skill routing.
- Treat the TOML files under
.codex/agentsas the only authored fleet. Files under.claude/agentsare generated from them bysync_claude_agents.py --write; never hand-edit a mirror. - Keep wave-1 agents read-only and leave branch switching, writes, approval, merge, deploy, secrets, and final decisions to the parent/governor.
- Keep global limits bounded at
max_threads = 6andmax_depth = 1unless a later review proves otherwise. - Route product-direction, founder-language, One/Kai/Nav, PKM, voice/action, and PR north-star lanes through the Founder Wiki North-Star Probe when material; repo/wiki divergence is
current_state_vs_north_star_drift. - Run agent validation, fleet audit, router smoke, mirror sync check, skill lint, and repo/docs governance checks after orchestration changes.
Handoff Rules
- Broad repo intake routes to
repo-context. - Generic skill-system authoring routes to
codex-skill-authoring. - CI, deploy, or runtime-governance follow-up routes to
repo-operations. - Future-state agent-lattice planning routes to
future-planner. - Domain implementation routes to the relevant owner skill.
Required Checks
python3 .codex/skills/codex-skill-authoring/scripts/truth_first_smoke.py
python3 .codex/skills/agent-orchestration-governance/scripts/agent_orchestration_check.py
python3 -m py_compile .codex/skills/agent-orchestration-governance/scripts/agent_orchestration_check.py .codex/skills/agent-orchestration-governance/scripts/delegation_router.py .codex/skills/agent-orchestration-governance/scripts/agent_fleet_audit.py .codex/skills/agent-orchestration-governance/scripts/agent_router_smoke.py .codex/skills/agent-orchestration-governance/scripts/subagent_budget.py .codex/skills/agent-orchestration-governance/scripts/sync_claude_agents.py
python3 .codex/skills/agent-orchestration-governance/scripts/agent_fleet_audit.py --text
python3 .codex/skills/agent-orchestration-governance/scripts/agent_router_smoke.py
./scripts/ci/repo-governance-check.sh
python3 .codex/skills/codex-skill-authoring/scripts/skill_lint.py
./bin/hushh codex audit
When not to use it
- →For broad repo scans that should start with `repo-context`
- →For generic skill creation or taxonomy work belonging to `codex-skill-authoring`
- →For domain implementation after the correct owner lane is clear
Limitations
- →Does not handle recursive multi-agent expansion beyond bounded defaults without evidence
- →Keeps global limits bounded at `max_threads = 6` and `max_depth = 1` unless reviewed
- →Wave-1 agents are kept read-only
How it compares
This skill provides a structured governance mechanism for multi-agent systems, unlike ad-hoc agent configuration.
Compared to similar skills
agent-orchestration-governance side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| agent-orchestration-governance (this skill) | 0 | 3mo | Review | Advanced |
| autonomous-agents | 10 | 6mo | No flags | Advanced |
| agent-goal-planner | 2 | 6mo | No flags | Advanced |
| planner | 1 | 6mo | No flags | Intermediate |
Try saying
Example prompts that trigger this skill in your AI assistant.
More by hushh-labs
View all by hushh-labs →You might also like
autonomous-agents
davila7
Autonomous agents are AI systems that can independently decompose goals, plan actions, execute tools, and self-correct without constant human guidance. The challenge isn't making them capable - it's making them reliable. Every extra decision multiplies failure probability. This skill covers agent loops (ReAct, Plan-Execute), goal decomposition, reflection patterns, and production reliability. Key insight: compounding error rates kill autonomous agents. A 95% success rate per step drops to 60% b
agent-goal-planner
ruvnet
Agent skill for goal-planner - invoke with $agent-goal-planner
planner
solatis
Interactive planning and execution for complex tasks. Use when user asks to use or invoke planner skill.
hierarchical-coordinator
MadAppGang
Prevent goal drift in long-running multi-agent workflows using a coordinator agent that validates outputs against original objectives at checkpoints. Use when orchestrating 3+ agents, multi-phase features, complex implementations, or any workflow where agents may lose sight of original requirements. Trigger keywords - "hierarchical", "coordinator", "anti-drift", "checkpoint", "validation", "goal-alignment", "decomposition", "phase-gate", "shared-state", "drift detection".
workflow-router
parcadei
Goal-based workflow orchestration - routes tasks to specialist agents based on user goals
loki-mode
davila7
Multi-agent autonomous startup system for Claude Code. Triggers on "Loki Mode". Orchestrates 100+ specialized agents across engineering, QA, DevOps, security, data/ML, business operations, marketing, HR, and customer success. Takes PRD to fully deployed, revenue-generating product with zero human intervention. Features Task tool for subagent dispatch, parallel code review with 3 specialized reviewers, severity-based issue triage, distributed task queue with dead letter handling, automatic deployment to cloud providers, A/B testing, customer feedback loops, incident response, circuit breakers, and self-healing. Handles rate limits via distributed state checkpoints and auto-resume with exponential backoff. Requires --dangerously-skip-permissions flag.