A Claude Code session reaches the platform's agents as subagents (Agent(subagent_type: "<slug>") on a committed .claude/agents/powernode/<slug>.md skeleton). The skeleton is a thin loader: it holds no domain knowledge. What the agent knows for THIS task, and which model and effort fit it, is decided by the platform's assembler (assemble_task) before the subagent is spawned.
Recall: search_knowledge tag guidance-claude-code-assembly.
- Call
assemble_taskwithtask,agent_slugandcontext: "claude_code". It returnsassembly_id,skills[](each with its prompt and verb list),learnings[],tier,model(haiku|sonnet|opus|fable),effort(low|medium|high|xhigh|max),floor_reasonanduncovered[]. - Spawn the subagent with the returned
modelandefforton the Agent tool, and putassembly_id: <id>in the delegating prompt. - The subagent's first move is
assemble_taskwith thatassembly_id: it gets the stored assembly back unchanged (idempotent), adopts the skills and learnings, loads only the verbs those skills name, does the work, and reports withrecord_agent_executionpassing the sameassembly_id. - The report links the assembly to the execution and writes one skill-usage record per selected skill, so the run credits the skills it used.
CLAUDE.md carries the one-line form of this rule. It also satisfies the Agent tool's requirement that effort is only set when the project says so.
- Tier is a floor, never an agent-wide pin. Core's roster tiers are concierge
standard, engineerstandard, reviewerreasoning; an extension-registered agent declares its own (e.g. the system extension's fleet operator,standard). The assembler raises the tier from two independent signals: the selected skills'hints(min_tier,effort) and verb risk (any gated, destructive or approval-parking verb in a selected skill's manifest raises toreasoning/high). The highest wins; nothing lowers below the roster tier. So a mutating fleet task runsopus/highand an inspect tasksonnet/medium. - Candidates stay inside the agent's domain ceiling (the concierge never gets platform-dev skills, the reviewer never gets front-door ones); a
requiresdependency of a selected skill rides along. - A Claude Code subagent never routes to a non-Claude provider:
context: "claude_code"returns a Claude model only. - A task no skill owns comes back with
uncovered[]and no skills: route it (route_task) or ask, do not improvise.
Hard enforcement is impossible client-side:
.claude/hooks/assembly-before-spawn.sh(PreToolUse onAgent) is ADVISORY: it warns when noassemble_taskfor that slug precedes the spawn in the transcript, or when the prompt carries noassembly_id..claude/hooks/session-guidance-inject.shprints the protocol line only when the connector listsassemble_task.- The server-side evidence is the assembly row (an
Ai::RoutingDecisionwithrequest_type: "task_assembly") linked to the execution byrecord_agent_execution.
- A subagent cannot spawn subagents. A delegated
platform-engineertherefore reports its diff and the delegating session (or the loop driver) runs the independent review; the review stays required. platform-reviewerhas no shell in phase 1 (no Bash, Edit or Write): it reads the diff and files with Read/Grep/Glob and the read-only code verbs, and names any check that needs a command (git, tests) in its report for the delegating session to run..claude/hooks/readonly-shell-guard.shwas not enough on its own: a frontmatter hook needs workspace trust and an unset$CLAUDE_PROJECT_DIRfails open (a non-blocking exit 127).- A deploy is never the engineer's: it is handed, through
route_task, to the agent that operates the running fleet (an extension-registered agent, e.g. the system extension's fleet operator). - A thin loader whose agent is not seeded on the connector stops and reports it to the driver without calling
record_agent_execution(which refuses a slug with no agent); the driver seeds the agents (see Deploying the roster) or uses a general-purpose agent for that task.