repair-spawn-intermediate
Generate a new intermediate task when no existing task produces a required output
- License 许可
- MIT
- 最近更新
- 2026-06-13
- Stars
- ★ 3
- Forks
- 1
01 前言 Front Matter
- name
- repair-spawn-intermediate
- tags
- [ "repair", "gap:missing-intermediate", "gap:blocker" ]
- context
- [ { "type": "gap", "fields": [ "gapKind", "taskId", "taskTitle", "inputPattern", "missingOutputs", "requiredByTask", "epicId", "suggestedUpstreamTask" ] }, { "type": "files", "label": "all-task-skills", "pattern": ".converge/epics/*/tasks/*/SKILL.md", "maxFiles": 30 }, { "type": "ai", "label": "task-plan", "tools": [ "Read", "Glob", "Grep" ], "prompt": "A task `{requiredByTask}` needs input files matching `{inputPattern}` but no existing task produces them.\nSearch all SKILL.md files in .converge/epics/ to confirm no task has these in its outputs. Then determine: 1. What kind of task should produce these files (e.g., generate HTML, compile data, screenshot) 2. What inputs the new task would need 3. Which epic it belongs to 4. A reasonable task ID (zero-padded, e.g., 002-generate-screenshots)\nReturn a concrete task definition plan in 5-10 lines.\n", "timeoutMs": 45000 }, { "cmd": "ls .converge/epics/{epicId}/tasks/ 2>/dev/null | tail -5", "type": "cmd", "label": "existing-tasks-in-epic" } ]
- description
- Generate a new intermediate task when no existing task produces a required output
- allowed-tools
- [ "Read", "Write", "Bash", "Glob", "Grep" ]
03 SKILL.md 正文
Repair: Spawn Intermediate Task
No existing task produces the required output. Create a new task definition.
Step 1: Read Context
- Read
repair-context/gap.md— what's missing and who needs it - Read
repair-context/task-plan.md— AI plan for the new task - Read
repair-context/existing-tasks-in-epic.txt— current tasks for numbering - Read
repair-context/all-task-skills.md— all existing tasks for reference - Read
repair-context/history.md(if exists) — DO NOT repeat failed approaches
Step 2: Create the Task
Based on the AI task plan, create a new SKILL.md file:
- Determine the task ID — use the next available number in the epic
- Create the directory:
.converge/epics/{epicId}/tasks/{taskId}/ - Write SKILL.md with:
name: the task IDinputs: what the task needs to readoutputs: the files that the blocked task needs (from the gap)checks:test -ffor each output- Body: clear instructions for what to generate
---
name: {taskId}
description: Generate {missing files} for {downstream task}
inputs:
- {source files}
outputs:
- {the missing files from the gap}
checks:
- id: output-exists
cmd: "test -f {output}"
description: "{output} exists"
allowed-tools:
- Read
- Write
- Bash
---
# {Task Title}
Generate the missing files needed by {downstream task}.
## Instructions
{Based on AI task plan}
Step 3: Verify
# Verify the SKILL.md is valid YAML
node -e "require('yaml').parse(require('fs').readFileSync('.converge/epics/{epicId}/tasks/{taskId}/SKILL.md','utf8').match(/^---\n([\s\S]*?)\n---/)[1])"
Important
- The new task will be picked up by the orchestrator in the next iteration
- Do NOT try to execute the task yourself — just create the definition
- Ensure outputs match EXACTLY what the blocked task needs as inputs
04 同仓库更多 Skills
hello-checker
Use this skill when asked to "verify", "check", or "validate" a file's existence and content. It provides a reliable workflow for confirming files meet their spec.
ask-skill
暂无简介
html-review-artifact
Framework-wide guidance for crafting polished human-review artifacts as markdown or HTML
converge-planning
Design a Converge playbook end-to-end. Use when the user says "plan a project", "design a playbook", "decompose into tasks", "scaffold a Converge workflow", "onboard to Converge", or asks how to write TASK.md / playbook.yml. Two entry points: plan a new playbook or update an existing one. Hand off to `/converge-control` after producing the playbook structure.
converge-control
Operate a Converge playbook end-to-end — start runs, monitor progress, narrow failures with `--select`, inspect runtime state, validate definitions, surgically reset subtrees, and stop stuck or orphaned runs. Use this skill whenever the user says "run my playbook", "execute this", "monitor the run", "converge failed", "the run is stuck", "retry failures", "babysit this playbook", "what's the status?", "show me the gantt", or asks how to use `converge run` / `--select` / `inspect` / `doctor` / `clean` / `stop`. Also use it for `HTTP 401` / `Invalid API key` failures on the first task — those are environment-not-playbook problems — and for any post-run cleanup work. Hand off to `converge-planning` if the user wants to design or restructure the playbook itself.
repair-taskts
Fix task.ts definition issues — Seed syntax, missing .build(), empty checks, import errors, TypeScript errors