capsule-driver
Use to drive a slice's phases to completion through in-session capsule subagents in the MAIN worktree — you are the super-orchestrator seat, spawning capsule-orchestrators that each charter ~3 phases and spawn their own planners and workers. No worktrees, no dispatch, no confinement. Use when you want a whole slice implemented under a tight context budget without handing the tree to confined workers. Not for confined or parallel-worktree execution — that is /dispatch.
How do I install this agent skill?
npx skills add https://github.com/davidlee/doctrine --skill capsule-driverIs this agent skill safe to install?
- Gen Agent Trust Hubpass
The skill acts as an orchestrator that manages multiple tiers of subagents to implement complex coding tasks. Its primary security surface is the ingestion of external plan files, which could lead to indirect prompt injection if those files contain malicious instructions.
- Socketpass
No alerts
- Snykpass
Risk: LOW · No issues
What does this agent skill do?
Capsule driver (the super-orchestrator seat)
Drive a slice's phases to completion through a tier of subagents in the main worktree. You hold the human's attention; everything below you is disposable context.
Announce at start: "Using the capsule-driver skill to drive SL-NNN through capsule orchestrators."
Why this exists
A slice's implementation costs far more context than one agent should hold. The tiering exists so that no single context exceeds ~250k tokens, where both cost and output quality degrade:
you (interactive, human attached)
└── capsule-orchestrator charters ~3 phases, verifies, commits
├── capsule-phase-planner one phase → runtime sheet
└── capsule-worker one phase → source delta
Your own discipline is to spend as little context as possible. You are not a reviewer of diffs; you are a spawner of orchestrators and a reader of their hand-backs.
Preconditions — check these before spawning anything
-
Session permission posture. Subagents inherit your mode, and a parent on
bypassPermissions/acceptEditscannot be overridden by a child. Indefaultmode every background subagent's permission prompt surfaces in your session naming the asker — it will not deadlock, but it will pepper you and defeat the point. Confirm the operator is running a posture they are happy to have inherited by the whole tree. -
Spawn depth. The tree needs 2 layers below you. The default (3) is enough, so nothing needs setting. Only a value set too low blocks: if
CLAUDE_CODE_MAX_SUBAGENT_SPAWN_DEPTHis set below 2, at the limitAgentis silently withheld — an orchestrator that cannot spawn will quietly do the work itself. Check withprintenv CLAUDE_CODE_MAX_SUBAGENT_SPAWN_DEPTH; unset or ≥ 2 passes. -
No competing writers. Everything runs in the main worktree with no confinement. Confirm no other agent is active in this tree before you start.
-
The slice is planned. Design locked,
plan.tomlauthored, phases exist. This skill implements a plan; it does not make one. -
Background tasks off. Export
CLAUDE_CODE_DISABLE_BACKGROUND_TASKS=1into the session env before you start. The whole tree is strictly serial —TaskOutputis withheld from every subagent, so no agent here can even read a running child — and the harness backgrounds by default, so every spawn becomes a completion notification that re-invokes its parent at that parent's full context. Without the switch the tiers below fall back to asking for foreground, which is model-mediated, not guaranteed.Set it out of band — per session, never in
settings.json. It is scoped to this seat's run, not to the repo: it also disablesrun_in_backgroundon Bash, which other work in this tree legitimately uses. Do not make it a project setting to save a step.
The loop
- Charter an orchestrator: a named slice and a contiguous run of phases,
typically 3. Spawn
capsule-orchestratorwith that charter, the slice id, and any standing constraint the human has set. Spawn it unnamed — never passname. A named spawn makes the orchestrator a teammate, and a teammate's children cannot hand back to it: their reports land at your seat ("the agent that spawned you is no longer running") instead. - Read its hand-back. Check the per-worker model choices carry real arguments — "the default" is not one. Check what it adapted in the plan.
- Reap it.
TaskStopthe orchestrator the moment you hold its hand-back. It is discharged; left standing it can be woken by a stray notification and pay a full turn at its accumulated context to report nothing. Confirm it is gone from/tasksbefore you charter the next one. - Charter the next orchestrator for the next run of phases. Always spawn a fresh one rather than resuming a tired one.
- Record friction as it is reported to you (
doctrine observation record). - Halt when phase implementation is complete. Do not begin audit.
What you escalate to the human
Subagents cannot ask — AskUserQuestion is withheld from every subagent, so an
orchestrator's only channel is its return value. You are the entire escalation
path. Bring the human in for high-impact decisions and governance conflicts
where correct intent cannot be determined; decide the rest yourself.
Boundaries
- Not
/dispatch. No worktrees, no forks, no confinement, no import step. Workers edit the real tree and orchestrators commit directly. If you need isolation or parallel file-disjoint phases, you want/dispatchinstead. - Do not implement. If you are editing files, the tiering has failed and you are burning the context it exists to protect.
- Do not audit.
/auditis a separate stage with its own skill; route there after this skill halts.
Related
/plan and /phase-plan produce what this consumes. /audit is what comes
after. /dispatch is the confined-worktree alternative to this skill.
How can the creator link this skill?
Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.
<a href="https://skillzs.dev/skills/davidlee/doctrine/capsule-driver">View capsule-driver on skillZs</a>