跳到正文
cccskills
免费在 GitHub 公开

sling-dispatch

Instruction dispatch pipeline for routing work items to available agents. Implements a 7-stage fetch-allocate-prepare-hook-store-launch-confirm pipeline with batch convoy mode, formula expansion, idempotent dispatch, and crash recovery.

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(3)

  • SKILL.md5.4 KB
  • references/gastown-origin.md6.2 KB
  • references/pipeline-implementation.md12.0 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Sling Dispatch

Instruction dispatch unit for the Gastown chipset. The sling routes a work item (bead) to an available polecat through a 7-stage pipeline. In hardware terms, the sling is the instruction fetch-decode-dispatch unit — it sets up the conditions for execution and confirms the dispatch was recorded, but does not execute work itself.

Activation Triggers

This skill activates when:

  • The mayor needs to assign work items to available polecats
  • A convoy of beads is ready for dispatch to a rig
  • A formula template needs expansion into ordered dispatch steps
  • A partial dispatch needs recovery after a crash or session restart

The 7-Stage Dispatch Pipeline

Every dispatch flows through seven stages in strict order. Each stage completes before the next begins. If any stage fails, the pipeline halts at that stage and the partial state is recoverable.

#StagePurpose
1FetchResolve bead ID to a work item; reject anything not in open status (idempotency).
2AllocateSelect an idle polecat from the pool, or spawn a new one if the pool has capacity.
3PrepareSet up the polecat's isolated workspace (worktree + execution context).
4HookSet work item status to hooked and bind it to the polecat. Idempotency boundary.
5StoreWrite dispatch arguments to durable state so a crash after this point is recoverable.
6LaunchStart the polecat's agent session and send assignment mail. GUPP activates immediately.
7ConfirmRecord the bead-to-agent assignment in convoy tracking and notify the witness.

Full TypeScript implementations live in references/pipeline-implementation.md.

Batch Mode

When more than batch_threshold beads (default: 3, configurable in chipset YAML at dispatch.batch_threshold) target the same rig, the sling automatically batches them into a convoy and dispatches in parallel. The batch detection groups beads by target rig; groups exceeding the threshold become convoys. Dispatches within a batch run in parallel up to max_parallel. Convoy progress updates after each individual dispatch confirms.

Formula Mode

When a formula (TOML template) is specified, the sling expands it into step-ordered beads before dispatching. Formulas define multi-step workflows as declarative sequences. Beads respect dependency ordering — steps with depends_on wait for their dependencies to reach done status before being dispatched. Steps without dependencies dispatch immediately.

Idempotency

The sling is idempotent. Re-dispatching an already-hooked bead is a no-op:

  • Stage 1 (Fetch) rejects any bead not in open status
  • A bead transitions to hooked in Stage 4 and never returns to open
  • Calling dispatchSingle(beadId) multiple times for the same bead has the same effect as calling it once
  • No side effects on repeated calls — the second call returns immediately at Stage 1

This makes the dispatch pipeline safe to retry after crashes, timeouts, or duplicate requests.

Crash Recovery

If the sling crashes mid-pipeline, the dispatch can be resumed from the durable state written in Stage 5 (Store). Records in the stored stage resume from Launch; records in the launched stage resume from Confirm; confirmed records are complete. All state updates use the write-then-rename pattern so partial writes are impossible.

Communication Protocol

The sling is invoked by the mayor, not by external events. The mayor decides what to dispatch and when; the sling handles how. On the outbound side, the sling sends durable mail to the polecat (work assignment) and the witness (dispatch confirmation), and writes the polecat's hook record in Stage 4. Inbound, it only receives dispatch and batch-dispatch mail from the mayor.

Boundary: What the Sling Does NOT Do

The sling dispatches; it does not execute work (the polecat does), decide what to dispatch (the mayor does), monitor agent progress (the witness does), resolve merge conflicts (the refinery does), or modify bead content. It fetches instructions, decodes them, assigns them to polecats, and confirms delivery.

Integration with Other Gastown Skills

Invoked by mayor-coordinator; targets polecat-worker via hook-persistence (Stage 4) and mail-async (Stages 6-7). Uses beads-state (StateManager) for work items and dispatch records. Emits confirmations to witness-observer. done-retirement processes the other end of the lifecycle.

References

Implementation detail moved to:

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Provides web accessibility best practices for semantic HTML, ARIA, keyboard navigation, color contrast, and screen reader patterns. Use when building UI components, reviewing accessibility, or when user mentions 'a11y', 'accessibility', 'ARIA', 'screen reader', 'keyboard navigation', 'WCAG'.

中文简介正在准备中。当前显示原文说明。

Tibsfox/gsd-skill-creator712026年7月20日 更新

Active listening techniques for effective communication. Covers attending behaviors, paraphrasing, reflective listening, clarifying questions, empathic response, barriers to listening, listening in conflict, and cross-cultural listening. Use when building listening skills, improving understanding in conversation, mediating disputes, or analyzing communication breakdowns.

中文简介正在准备中。当前显示原文说明。

Tibsfox/gsd-skill-creator712026年7月20日 更新

Adversarial spec-compliance PR review — cross-references diffs against approved specs, verifies runtime claims against source, detects competing PRs, audits scope/convention compliance. Use before merging.

中文简介正在准备中。当前显示原文说明。

Tibsfox/gsd-skill-creator712026年7月20日 更新

Provides best practices for AI agent orchestration including MCP servers, A2A protocol, multi-agent coordination, and swarm architectures. Use when designing agent systems, configuring MCP servers, setting up agent teams, or when user mentions 'MCP', 'A2A', 'agent orchestration', 'multi-agent', 'swarm', 'agent team', 'LangGraph', 'CrewAI', 'AutoGen'.

中文简介正在准备中。当前显示原文说明。

Tibsfox/gsd-skill-creator712026年7月20日 更新

Symbolic manipulation, equation solving, and algebraic structures for mathematical reasoning. Covers distributive law, factoring, completing the square, linear through polynomial equation solving, systems of equations (substitution, elimination, Gaussian elimination, matrix methods), algebraic structures (groups, rings, fields), modular arithmetic, polynomial theory, and inequalities. Use when solving equations, simplifying expressions, working with algebraic structures, or performing symbolic manipulation.

中文简介正在准备中。当前显示原文说明。

Tibsfox/gsd-skill-creator712026年7月20日 更新

Understanding how algorithmic systems shape what users see, know, and do -- from recommendation feeds to search ranking to credit scoring to hiring software. Covers the mechanics of recommendation systems, algorithmic bias and its sources, personalization's effects on information diets, opacity and accountability, AI limitations (hallucination, confident wrongness), and the human-in-the-loop question. Use when a learner needs to think critically about why particular content reached them.

中文简介正在准备中。当前显示原文说明。

Tibsfox/gsd-skill-creator712026年7月20日 更新

查看 Tibsfox 的全部技能

报告此技能的问题