Prompt file imported from materialgram/materialgram (.claude/commands/continue.md). Fill in {{arguments}} before use. Copyright stays with the author.

Continue AI Work

Name: continue
Author: materialgram

Read .claude/ai-workflow-adapter.md and .agents/skills/continue/SKILL.md completely. Follow the shared skill with the Claude adapter's delegation and text-handling substitutions. At startup choose exactly one shared-skill mode and record its frozen task ids: resume only the active task, drain only the existing queue snapshot, or, when both are empty, process one inbox transaction and drain only its routed tasks. Never add unrelated tasks observed later in the run. Append and implement only new follow-up tasks routed from results produced by this invocation, including their transitive discovered follow-ups. Continue until that frozen-and-derived batch reaches the scheduler's normal stop condition or a global hard stop.

Every inbox worker, performer, and discovered-task routing Agent must be told in its initial prompt to read .claude/ai-workflow-adapter.md completely before the applicable shared skill or reference. Use the Agent tool for those workers; do not start Claude subprocesses through Bash.

Arguments, when present, are natural-language hints for new shared work; their wording decides whether they merely prioritize or strictly restrict selection:

{{arguments}}

# Continue AI Work Read `.claude/ai-workflow-adapter.md` and `.agents/skills/continue/SKILL.md` completely. Follow the shared skill with the Claude adapter's delegation and text-handling substitutions. At startup choose exactly one shared-skill mode and record its frozen task ids: resume only the active task, drain only the existing queue snapshot, or, when both are empty, process one inbox transaction and drain only its routed tasks. Never add unrelated tasks observed later in the run. Append and implement only new follow-up tasks routed from results produced by this invocation, including their transitive discovered follow-ups. Continue until that frozen-and-derived batch reaches the scheduler's normal stop condition or a global hard stop. Every inbox worker, performer, and discovered-task routing Agent must be told in its initial prompt to read `.claude/ai-workflow-adapter.md` completely before the applicable shared skill or reference. Use the Agent tool for those workers; do not start Claude subprocesses through Bash. Arguments, when present, are natural-language hints for new shared work; their wording decides whether they merely prioritize or strictly restrict selection: ```text {{arguments}} ```

{ "prompt": "# Continue AI Work\n\nRead `.claude/ai-workflow-adapter.md` and\n`.agents/skills/continue/SKILL.md` completely. Follow the shared skill with the\nClaude adapter's delegation and text-handling substitutions. At startup choose\nexactly one shared-skill mode and record its frozen task ids: resume only the\nactive task, drain only the existing queue snapshot, or, when both are empty,\nprocess one inbox transaction and drain only its routed tasks. Never add\nunrelated tasks observed later in the run. Append and implement only new\nfollow-up tasks routed from results produced by this invocation, including\ntheir transitive discovered follow-ups. Continue until that frozen-and-derived\nbatch reaches the scheduler's normal stop condition or a global hard stop.\n\nEvery inbox worker, performer, and discovered-task routing Agent must be told\nin its initial prompt to read `.claude/ai-workflow-adapter.md` completely before\nthe applicable shared skill or reference. Use the Agent tool for those workers;\ndo not start Claude subprocesses through Bash.\n\nArguments, when present, are natural-language hints for new shared work; their\nwording decides whether they merely prioritize or strictly restrict selection:\n\n```text\n{{arguments}}\n```", "variables": [ "arguments" ], "notes": "Continue work from the shared ai-tdesktop queue Tools: Read, Write, Edit, Glob, Grep, Bash, Agent, AskUserQuestion, TodoWrite.", "metadata": { "source": { "provider": "github-claude-commands", "repository": "https://github.com/materialgram/materialgram", "path": ".claude/commands/continue.md", "ref": "bccbc4336b5dbff70f16c5fc5ad1cf1d27972c72", "url": "https://github.com/materialgram/materialgram/blob/bccbc4336b5dbff70f16c5fc5ad1cf1d27972c72/.claude/commands/continue.md", "key": "materialgram/materialgram/.claude/commands/continue.md" } } }

continue

About

Continue AI Work

Use it

Manifest

Reviews