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

evals/bench/tasks.json

View on GitHub
{
  "version": 1,
  "tasks": [
    {
      "id": "T1",
      "commit": "ab9ecaa8bbc6ec389d3a7ac8672a8df6cf5cbdef",
      "title": "file-tidy: harden Linux trash for byte names, other mounts and relative XDG_DATA_HOME",
      "prompt": "file-tidy's Linux (freedesktop) trash support in `skills/file-tidy/scripts/file_tidy.py` has three bugs.\n\n1. `write_trashinfo(dest, original)` breaks on file names that are not valid UTF-8 (e.g. a POSIX name containing byte 0xE9). The `Path=` line must percent-encode the raw file-name bytes, keeping `/` unescaped: for original `os.fsdecode(b\"/srv/caf\\xe9 1.txt\")` the file must contain `Path=/srv/caf%E9%201.txt`.\n\n2. `trash_dir()` honours a relative `XDG_DATA_HOME`. The XDG spec says a relative value must be ignored; on Linux with `XDG_DATA_HOME=relative/share` it must return `~/.local/share/Trash/files`.\n\n3. Trashing an item that lives on another mount (e.g. a USB drive) silently copies it into the home trash. Add a module-level function `on_trash_volume(src: Path) -> bool` that reports whether `src` is on the same device as the home trash (use the nearest existing ancestor if the trash folder does not exist yet). When planning an apply (`apply <plan> --apply`), a `trash` item on a freedesktop system (`freedesktop_trash()` true) whose source is not on the trash volume must abort with `SystemExit` whose message contains `use move instead`, before anything moves.\n\nTests run with `python3 -m unittest discover -s skills/file-tidy/scripts -p 'test_*.py'`; add tests for each case. Nice to have: mention in the Undo section of `skills/file-tidy/SKILL.md` that trash refuses items on another mount.",
      "hidden_tests": ["skills/file-tidy/scripts/test_file_tidy.py"],
      "accept": [
        "python3 -m unittest discover -s skills/file-tidy/scripts -p 'test_*.py'"
      ],
      "needs_node_modules": false,
      "parts": "One module: file_tidy.py plus its test file; the SKILL.md line is a one-sentence doc touch, not worth a separate worker.",
      "verified": {
        "parent_fails": true,
        "commit_passes": true
      }
    },
    {
      "id": "T2",
      "commit": "09d3be909fa5db65cf1f1dcaa9a88c43ae1cae93",
      "title": "worktree-tidy: keep ignored secrets and unpushed branch commits",
      "prompt": "`skills/worktree-tidy/scripts/worktree_tidy.py` can lose work: `git worktree remove` silently deletes ignored files (e.g. `.env`), and `--delete-branch` runs `branch -D` after a merged request even if local commits were added later.\n\nRequired:\n- `dirty_paths(path)` makes exactly one `git --no-optional-locks status ... --ignored=matching` call and returns `(dirty, ignored)` or None. Ignored paths are fine when regenerable: any segment in module set `REGENERABLE_DIRS` (node_modules, build, generated, ... and similar), a prefix in `REGENERABLE_PREFIXES` (`.husky/_/`), or `*.log`. Secrets (`.env`, `.env.*`, `id_rsa*`, `*.pem`, `*.key`, ...) are never regenerable, even inside collapsed build/generated folders (walk them, but not node_modules). Other ignored paths add Blocker `ignored`. Scan uses at most 6 subprocess calls per worktree.\n- `add_regenerable(list)` extends those sets (entries with `/` are prefixes); non-list raises SystemExit; config key `regenerable`.\n- Each row adds `shortHead`, `subject`, `dirtyFiles` (max 10, rename targets included), `ignored` (count), `ignoredFiles` (max 10), `branchSaved` (branch commits all on another branch/tag/remote ref), `unpushed` (HEAD commits on no remote ref; null without remote refs). Missing folders: empty/0/False/None.\n- `branchDeletable` and dry-run `wouldDeleteBranch` also require `branchSaved`.\n- `remove(paths, days, delete_branch, use_orca, yes, allow_recent=False)` and CLI `--allow-recent`: allow removal when blockers are exactly `[\"recent\"]`.\n- Right before removing, re-check status; on change error `not removable: changed since it was checked`. Before deleting a branch re-check it; if unsaved keep it with `branchError` `branch has commits that exist on no other ref; kept`.\n\nDocument these in SKILL.md (nice-to-have).",
      "hidden_tests": ["skills/worktree-tidy/scripts/test_worktree_tidy.py"],
      "accept": [
        "python3 -m unittest discover -s skills/worktree-tidy/scripts -p 'test_*.py'"
      ],
      "needs_node_modules": false,
      "parts": "Mostly one module (worktree_tidy.py); SKILL.md and references/glossary.md docs can go to a second worker, but the code changes in dirty_paths/assess/remove are interdependent and should not be split.",
      "verified": {
        "parent_fails": true,
        "commit_passes": true
      }
    },
    {
      "id": "T3",
      "commit": "40087abbbbbbd331ff2bd43e63a75bdc791b8d97",
      "title": "validate-skills: enforce the 1024-character, tag-free description and fit figma-mcp in it",
      "prompt": "`scripts/validate-skills.mjs` only checks that a skill `description` is at least 20 characters. The Agent Skills spec also requires at most 1024 characters and no XML tags, and `skills/figma-mcp/SKILL.md` currently violates both (it is over 1024 characters and contains `<url>`).\n\nRequired:\n- Export `DESCRIPTION_MAX_LENGTH` (1024) and `descriptionProblems(rawDescription)` from `scripts/validate-skills.mjs`. It trims the input, counts characters (not bytes), and returns an array of problem strings (empty when fine). Keep the existing too-short message (contains `过短`). For an over-long description the message must include the actual length as `<n> 字符` (e.g. `1025 字符`) and the fix hint `删掉重复的触发词`. For XML-like tags the message must list the tags found (e.g. `<url>`); plain comparisons such as `count < 30 and > 5` are not tags.\n- `main()` must report each problem per skill, and importing the module must not run validation (only run `main()` when the file is executed directly).\n- Shorten the `skills/figma-mcp/SKILL.md` description to at most 1024 characters with no angle-bracket tags, keeping its meaning; bump its `metadata.version` patch.\n- `pnpm skills:validate` must pass.\n\nNice to have: add `./scripts/__tests__/validate-skills.test.mjs` to the root `test:scripts` script and regenerate the skills index with `pnpm skills:index`.",
      "hidden_tests": ["scripts/__tests__/validate-skills.test.mjs"],
      "accept": [
        "node --experimental-strip-types --test scripts/__tests__/validate-skills.test.mjs",
        "pnpm skills:validate"
      ],
      "needs_node_modules": true,
      "parts": "Two independent file sets: scripts/validate-skills.mjs (+ package.json test:scripts) and skills/figma-mcp/SKILL.md description rewrite; they only meet at the final pnpm skills:validate.",
      "verified": {
        "parent_fails": true,
        "commit_passes": true
      }
    }
  ]
}
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