From agent-almanac
Scales distributed systems and organizations using colony budding, role differentiation, and architectural transitions to maintain coordination as agent counts grow past initial capacity.
How this skill is triggered — by the user, by Claude, or both
Slash command
/agent-almanac:scale-colonyThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Scale distributed systems, teams, or organizations through colony budding (splitting), role differentiation (age polyethism), and growth-triggered architectural transitions — maintaining coordination quality as the colony grows beyond its initial design capacity.
Scale distributed systems, teams, or organizations through colony budding (splitting), role differentiation (age polyethism), and growth-triggered architectural transitions — maintaining coordination quality as the colony grows beyond its initial design capacity.
Identify which scaling phase the colony is in to apply appropriate strategies.
Colony Growth Phases:
┌───────────┬──────────────┬───────────────────────────────────────────┐
│ Phase │ Size Range │ Characteristics │
├───────────┼──────────────┼───────────────────────────────────────────┤
│ Founding │ 1-7 agents │ Everyone does everything, direct comms, │
│ │ │ implicit coordination, high agility │
├───────────┼──────────────┼───────────────────────────────────────────┤
│ Growth │ 8-30 agents │ Roles emerge, some specialization, comms │
│ │ │ overhead increases, need for structure │
├───────────┼──────────────┼───────────────────────────────────────────┤
│ Maturity │ 30-100 agents│ Formal roles, layered coordination, │
│ │ │ sub-groups form, inter-group coordination │
├───────────┼──────────────┼───────────────────────────────────────────┤
│ Fission │ 100+ agents │ Colony too large for single coordination │
│ │ │ framework, must bud into sub-colonies │
└───────────┴──────────────┴───────────────────────────────────────────┘
Expected: Clear identification of the current growth phase and the specific stress signals indicating the colony is approaching or has crossed a phase boundary.
On failure: If the phase isn't clear, measure three concrete metrics: communication volume per agent, decision latency, and coordination failure rate. Plot them over time. The inflection points reveal phase transitions. If metrics aren't available, the colony is likely in the Founding phase (where metrics aren't yet needed).
Introduce progressive specialization where agents take on different roles based on experience and colony needs.
forage-resources)defend-colony)Expected: A role structure where agents naturally progress from simple to complex responsibilities, with the colony's role distribution reflecting its current needs and phase.
On failure: If role differentiation creates rigid silos, increase cross-training requirements and rotation frequency. If newcomers struggle to progress, the mentoring system is insufficient — pair each newcomer with a specialist for their first N tasks. If too many agents cluster in one role, the transition triggers are miscalibrated — adjust thresholds based on colony-wide role demand.
Adapt the coordination mechanisms from coordinate-swarm to handle increased colony size.
Expected: A layered coordination structure where communication overhead grows logarithmically (not linearly) with colony size. Local coordination is fast and direct; colony-wide coordination is slower but still functional.
On failure: If coordination layers create information bottlenecks (communicators become overloaded), add redundant communicators or reduce the relay frequency. If layers create isolation (squads don't know what other squads are doing), increase the inter-layer signal frequency or create cross-squad liaison roles.
Split the colony into semi-autonomous sub-colonies when it exceeds single-coordination capacity.
build-consensus)Expected: Two or more viable daughter colonies, each operating semi-autonomously with their own coordination, connected by lightweight inter-colony interfaces.
On failure: If daughter colonies are too small to be viable, the fission was premature — remerge and try again at a larger size. If inter-colony coordination becomes as heavy as pre-fission single-colony coordination, the split lines were wrong — the colonies are too interdependent. Re-draw boundaries along natural independence lines.
Continuously assess whether the current structure matches the colony's size and needs.
adapt-architecture)Expected: A colony that monitors its own scaling health and proactively adapts its structure before scaling stress becomes scaling failure.
On failure: If scaling health metrics are not available, the colony lacks observability — build measurement before building more structure. If metrics show problems but the colony can't adapt, the resistance is cultural, not technical — address the human factors (fear of change, ownership attachment, trust deficits) before restructuring.
coordinate-swarm — foundational coordination patterns that this skill scalesforage-resources — foraging scales differently than production; role differentiation affects scout allocationbuild-consensus — consensus mechanisms must adapt for larger groupsdefend-colony — defense must scale with the colonyadapt-architecture — morphic skill for structural transformation, triggered by growth pressureplan-capacity — capacity planning for growth projectionsconduct-retrospective — retrospectives help identify scaling stress before it becomes failurenpx claudepluginhub pjt222/agent-almanacCoordinates distributed agents via stigmergy, local rules, and quorum sensing to achieve emergent collective behavior without centralized control. Use for scaling multi-agent systems, self-organizing teams, or event-driven architectures.
Coordinates multi-agent swarms with queen-led architecture, Byzantine consensus, and shared persistent memory. Use for orchestrating agent teams under a lead coordinator or building resilient cross-agent collaboration.
Assesses team structure against Team Topologies, evaluating cognitive load, interaction modes, Conway's Law alignment, and fracture planes for optimized software delivery.