Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/friedbotstudio/baseline/swarm-dispatchnpx skills add friedbotstudio/baseline --skill swarm-dispatchgit clone --depth 1 https://github.com/friedbotstudio/baselineWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/friedbotstudio/baseline/swarm-dispatch)<a href="https://agentmods.dev/skills/friedbotstudio/baseline/swarm-dispatch"><img src="https://agentmods.dev/badge/skills/friedbotstudio/baseline/swarm-dispatch.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00084 | $0.03640 |
| Opus 5 | $0.00042 | $0.01820 |
| Sonnet 5 | $0.00017 | $0.00728 |
| Haiku 4.5 | $0.00008 | $0.00364 |
Grade A, and why
swarm-dispatch scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 236 lines — stays where its author put it; the contents beside it link to each section on GitHub.
swarm-dispatch — wave runner with worktree isolation
Invoked after /swarm-plan + /approve-swarm. The architecture is the user's principle made concrete:
Main context decides. Workers execute.
Per task, before dispatch, you (main context) produce two recipes:
- The scenario recipe — exactly which failing tests the worker should write.
- The implementation contract — exactly which source files the worker may touch and what behavior they must implement.
The worker's prompt contains both recipes verbatim. The worker invokes Skill(scenario) then Skill(implement) and reports JSON. It makes no design decisions.
Isolation modes
Read project.json → swarm.isolation (default "auto"):
"auto"→ chooseworktreeif the project root is inside a git repo (git rev-parse --is-inside-work-treesucceeds), elseshared."worktree"→ require a git repo; bail if absent."shared"→ never use worktrees; rely onswarm_boundary_guardfor runtime enforcement.
Default path is worktree. The rest of this document describes that mode. The shared fallback is at the end.
Worktree mode = single-wave only (D1). The worktree base commit is chosen by the Agent tool's
isolation:"worktree", not by baseline (observed forking from a stale ref on the first real swarm run), andswarm_merge.mjsapplies wave output without committing — so a multi-wave worktree plan's wave-N+1 worktrees branch from a base lacking wave-N output. Baseline cannot fix this, only refuse it. Before dispatch, callassertWorktreeWaveSafety({isolation, waves, baselineRef, worktreeBase})from.claude/skills/swarm-dispatch/worktree-safety.mjs(worktreeBase =git -C <wt> merge-base HEAD <wt-HEAD>once a worktree exists). Onok:false(multi-wave under worktree, orbaseline_ref≠ the worktree's real merge-base), abort and steer the plan to shared isolation — do not run the wave. Multi-wave plans belong insharedmode.
Prereqs (worktree mode)
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 236 lines · 84 tokens per session scan A a3e38191fd4c
swarm-dispatch is a skill published in the GitHub repository friedbotstudio/baseline (14 stars, last pushed yesterday), licensed Apache-2.0. It adds 84 tokens to every session and 3,640 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
dispatching-parallel-agents
The parallel fan-out primitive. Routed to by any skill or command that splits work across independent units and dispatches an agent per unit — subagent-driven-development, /sprint, parallel /review. It owns the dispatch/collect/funnel discipline: bound concurrency, isolate units, collect every result, dedupe overlap…
finishing-a-development-branch
The terminal step of /feature and /sprint. Routed to once commit-gate has cleared, to decide the branch's fate — merge via PR, open a PR, or discard. Direct merge to the default branch is forbidden; every change lands through a PR. Under /sprint the skill auto-selects "open PR" and surfaces the merge decision to the…
ca-chore
Sanctioned lane for non-behavioral work — docs-only edits, dependency bumps, reverts. Type-scaled gates; no TDD demanded of prose.
ca-metrics
Read-only 3-metric governance glance — override rate, small-lane rate, sprint low-confidence ratio — each with a trend arrow vs. the prior 20-commit window.
ca-task
The sanctioned task-board mutator — add a queued task, start one (flips to in-progress and stamps the date, minting a dotted ID on pick-up), or mark an in-progress task done. The only blessed write to open-tasks.md.
ca-audit
Assemble the governance record for a range — commits, overrides, ADRs, sprint auto-decisions, open questions, checkpoint findings — into one dated audit packet. Read-only.