agentleFS
hushh-labshushh-labs/hushh-research/.agents/skills/agent-orchestration-governance/SKILL.md

agent-orchestration-governance

Skill40 starsChanged 4 days agoApache-2.0View on GitHub

Use when changing repository engineering agents, subagent concurrency or depth, delegation policy, or handoff verification rules in hushh-research.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 5 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Avoid overlap with codex-skill-authoring for generic skill taxonomy and repo-context for broad repository intake.

    line 12
  2. Distinctive:

    Verify that a custom agent is justified.

    line 64
  3. Distinctive:

    Keep the fleet at the curated sweet spot: broad read-only evidence lanes, not one agent per skill.

    line 65
  4. Distinctive:

    Keep custom-agent Markdown files thin: role, sandbox, nicknames, root craft/bacterial-architecture inheritance hook, concise instructions, …

    line 67
  5. Very specific:

    Treat .agents/subagents/*.md as the only authored fleet.

    line 68
  6. Distinctive:

    Keep wave-1 agents read-only and leave branch switching, writes, approval, merge, deploy, secrets, and final decisions to the …

    line 69
  7. Very specific:

    Keep global limits bounded at max_threads = 6 and max_depth = 1 unless a later review proves otherwise.

    line 71
  8. Distinctive:

    Run agent validation, fleet audit, router smoke, canonical-source check, skill lint, and repo/docs governance checks after orchestration …

    line 73

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Hussh Agent Orchestration Governance Skill

Purpose and Trigger

  • Primary scope: agent-orchestration-governance-intake
  • Trigger on repo-scoped custom agents, canonical host configuration limits, delegation policy, child handoff contracts, or workflow orchestration changes.
  • Avoid overlap with codex-skill-authoring for generic skill taxonomy and repo-context for broad repository intake.

Coverage and Ownership

  • Role: owner
  • Owner family: agent-orchestration-governance

Owned repo surfaces:

  1. .agents/subagents
  2. .agents/config.json
  3. .agents/skills/agent-orchestration-governance
  4. .agents/workflows/agent-orchestration-governance
  5. AGENTS.md
  6. docs/operations/README.md
  7. docs/operations/coding-agent-mcp.md
  8. .agents/routing
  9. .agents/instructions.md
  10. .agents/README.md

Non-owned surfaces:

  1. repo-context
  2. codex-skill-authoring
  3. repo-operations
  4. future-planner

Do Use

  1. Adding or tightening repo-scoped custom agents.
  2. Changing subagent concurrency/depth limits or delegation policy.
  3. Defining authority boundaries, parent-only actions, or child handoff shape.
  4. Keeping agents thin and routed through existing owner/spoke skills.

Do Not Use

  1. Broad repo scans that should start with repo-context.
  2. Generic skill creation or taxonomy work that belongs to codex-skill-authoring.
  3. Domain implementation after the correct owner lane is clear.
  4. Recursive multi-agent expansion beyond bounded defaults without evidence.

Read First

  1. AGENTS.md
  2. .agents/skills/agent-orchestration-governance/references/delegation-contract.md
  3. .agents/skills/codex-skill-authoring/references/skill-contract.md
  4. .agents/skills/codex-skill-authoring/references/truth-first-operating-kernel.md
  5. docs/operations/coding-agent-mcp.md
  6. .agents/skills/agent-orchestration-governance/references/anti-rationalization.md

Workflow

  1. Verify that a custom agent is justified; prefer skills/workflows when role specialization is not needed.
  2. Keep the fleet at the curated sweet spot: broad read-only evidence lanes, not one agent per skill.
  3. Preserve the repo-wide delegation checkpoint and truth-first handoff shape in AGENTS.md and delegation-contract.md.
  4. Keep custom-agent Markdown files thin: role, sandbox, nicknames, root craft/bacterial-architecture inheritance hook, concise instructions, and skill routing.
  5. Treat .agents/subagents/*.md as the only authored fleet; no host projections or copies are created.
  6. Keep wave-1 agents read-only and leave branch switching, writes, approval, merge, deploy, secrets, and final decisions to the parent/governor.
  7. Reject vendor directories and per-host adapters; read .agents through root AGENTS.md. Release finished child sessions through supported host controls and verify closure.
  8. Keep global limits bounded at max_threads = 6 and max_depth = 1 unless a later review proves otherwise.
  9. 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.
  10. Run agent validation, fleet audit, router smoke, canonical-source check, skill lint, and repo/docs governance checks after orchestration changes.

Handoff Rules

  1. Broad repo intake routes to repo-context.
  2. Generic skill-system authoring routes to codex-skill-authoring.
  3. CI, deploy, or runtime-governance follow-up routes to repo-operations.
  4. Future-state agent-lattice planning routes to future-planner.
  5. Domain implementation routes to the relevant owner skill.

Required Checks

python3 .agents/skills/codex-skill-authoring/scripts/truth_first_smoke.py
python3 .agents/skills/agent-orchestration-governance/scripts/agent_orchestration_check.py
python3 -m py_compile .agents/skills/agent-orchestration-governance/scripts/agent_orchestration_check.py .agents/skills/agent-orchestration-governance/scripts/delegation_router.py .agents/skills/agent-orchestration-governance/scripts/agent_fleet_audit.py .agents/skills/agent-orchestration-governance/scripts/agent_router_smoke.py .agents/skills/agent-orchestration-governance/scripts/subagent_budget.py
python3 .agents/skills/agent-orchestration-governance/scripts/agent_fleet_audit.py --text
python3 .agents/skills/agent-orchestration-governance/scripts/agent_router_smoke.py
./scripts/ci/repo-governance-check.sh
python3 .agents/skills/codex-skill-authoring/scripts/skill_lint.py
./bin/hushh agents audit

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.