Skip to content

Adonis0123/adonis-skills / skills/agent-dispatch

agent-dispatch

Flagged

Choose the lightest agent setup for a coding, review, drafting or design-research task: do it yourself, fan out to 2-3 workers with independent parts across the Claude Code, Codex, Grok Build, Cursor or other CLIs and accounts on this machine, run a scoped architecture loop, or send one brief to independent researchers; pick a cross-model reviewer, test budget and cleanup; track dispatched work, external waits and user decisions on a Run Board. Use when work may need delegation, or the user asks to dispatch, fan out, parallelize, run several agents or accounts, get another AI's review or independent design opinions, or asks how dispatched work is going. Chinese triggers: 派活、并行、要不要多开几个 agent、分给多个 Claude 账号 / codex / grok、找另一个 AI 审、对比优秀设计、多方意见、进度看板、派出的任务进行到哪了。Not for unrelated read-only questions, trivial edits, research with no delegation, a single review of an existing diff (agentic-review-handoff), or the machine-wide session list (sideby sessions). Inside another named workflow, only size its workers.

FileKindSize
GLOSSARY.mdmarkdown2.1 KB
SKILL.mdskill md27.8 KB
evals/README.mdmarkdown7.6 KB
evals/README.zh-CN.mdmarkdown7.5 KB
evals/bench/compare.pyscript5.3 KB
evals/bench/run_bench.pyscript12.3 KB
evals/bench/tasks.jsonconfig6.6 KB
evals/bench/test_bench.pyscript4.4 KB
evals/build_plugin_eval.pyscript11.3 KB
evals/evals.jsonconfig31.4 KB
evals/results.mdmarkdown15.8 KB
evals/test_eval_harness.pyscript3.5 KB
evals/trigger-eval-set.jsonconfig1.6 KB
fixtures/probe-full.jsonconfig1.6 KB
fixtures/probe-single.jsonconfig690 B
fixtures/run-midrun/B9.jsonlother191 B
fixtures/run-midrun/deploy-log.txtmarkdown42 B
fixtures/run-midrun/run.jsonconfig1.8 KB
fixtures/run-sessions/B9.jsonlother191 B
fixtures/run-sessions/deploy-log.txtmarkdown42 B
fixtures/run-sessions/ended.jsonlother95 B
fixtures/run-sessions/run.jsonconfig2.1 KB
fixtures/run-sessions/sessions.jsonconfig804 B
fixtures/session-index.jsonconfig401 B
fixtures/sideby-stub/sessions.jsonconfig1.5 KB
fixtures/sideby-stub/sidebyexecutablescript4.1 KB
references/accounts.mdmarkdown10.4 KB
references/context-budget.mdmarkdown2.9 KB
references/fallbacks.mdmarkdown4.7 KB
references/launch.mdmarkdown16.1 KB
references/measure.mdmarkdown3.3 KB
references/research-brief.mdmarkdown1.3 KB
references/run-board.mdmarkdown48.7 KB
references/task-spec.mdmarkdown5.0 KB
scripts/board.cssother12.6 KB
scripts/board_interaction.jsscript3.7 KB
scripts/board_markdown.jsscript2.9 KB
scripts/board_render.pyscript17.2 KB
scripts/dispatch_board.pyscript95.3 KB
scripts/isolated_case.pyscript1.6 KB
scripts/probe_agents.pyscript12.9 KB
scripts/sideby_cli.pyscript3.2 KB
scripts/sideby_sessions.pyscript3.3 KB
scripts/test_billing_and_quota.pyscript10.9 KB
scripts/test_board_interaction.jsscript2.1 KB
scripts/test_board_markdown.jsscript2.9 KB
scripts/test_board_reload_state.jsscript2.7 KB
scripts/test_bug_regressions.pyscript4.7 KB
scripts/test_dispatch_board.pyscript68.9 KB
scripts/test_harness_limits.pyscript11.7 KB
scripts/test_lane_states.pyscript7.5 KB
scripts/test_probe_agents.pyscript19.2 KB
scripts/test_review_fixes.pyscript6.7 KB
scripts/test_rounds_and_watch.pyscript7.3 KB

references/context-budget.md

View on GitHub
# Keep the Coordinator's context small

On one measured Run (n=1; re-measure with `measure.md`) the Coordinator was about half of all spend (context grew to ~600k tokens; screenshots were the largest item), because every token it reads is re-read on every later turn.

1. **The ledger is the only state.** Do not keep Worker details in the conversation; re-read with `show --line` or `show` (short text).
2. **Never Read the Board HTML, `run.json`, Worker `.jsonl` streams or full reports.** Changing files are worse than large ones: Claude Code injects a file's diff into the conversation each time a file it read changes. The Board lives outside the repository and is rewritten only at milestones, but reading it still costs the whole page.
3. **Workers return a summary of at most 10 lines plus paths** (`task-spec.md` Return). Register them with `lane set --summary-file <summary> --result-file <report>`. Read the whole summary, not only its first sentence: its first line is for the user, the rest lists the Worker's decisions, deviations from the spec and UNVERIFIED items, which you need before integrating. Do not Read the full report; when one of its details matters, give the path to a subagent and take back its text answer.
4. **Screenshots are judged by a subagent that returns text**; look at the final one yourself only when the user will see it.
5. **Round budget: the first run plus up to 2 fix rounds** (SKILL.md "Dispatching workers"); each fix round's review checks only the previous round's findings. Mechanical reviews (copy, formatting) go to a cheaper model; logic and security get the cross-model reviewer.
6. **A shared contract lands before the parts that use it fan out** (SKILL.md tier card). Two merges in the measured Run broke because one Lane added a required field another Lane relied on.
7. **A fix round resumes the same Worker session** (`launch.md` § Fix rounds).
8. **Let the host compact automatically**; never ask the user to run `/compact` (they run no commands). After a compaction, re-read the Run with `show` before acting: the ledger, not the summary, is the state. A SessionStart hook with matcher `compact` can inject "read <ledger> with show".
9. **Stop processes by pid, never by a command-line pattern.**
10. **A Worker's context is capped too.** `show` flags a running Lane past `--context-handoff-tokens` (default 150k). Let it finish its step and write its progress file (`task-spec.md` § Progress), then start a fresh session on the same spec plus that file. Handing off from a progress file wastes one short re-read; continuing re-reads the whole long context on every turn, and a session that dies at a limit without one loses the work.
11. **Spend is counted by billing.** Subscription Lanes cost quota, not money; API Lanes cost money. Quota and budget checks run only before a Lane or fix round starts and never stop running work (SKILL.md "Dispatching workers").
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 322kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 322kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 322kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 198kUpdated Claude CodeCodexVerifiedInstall