Agtx Task Sweep
fynnfluegge/agtx
Breaks a conversation's results into feature-level tasks and pushes them to the agtx kanban board, where each task gets its own worktree and agent session.
Explicitly spawn a standalone child session under the current vlx-term session, passing the task in as its first message (mirrors spawntask).
Install Vspawn
$ npx skills add vlinx-io/VelaTerm --skill vspawn -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install vlinx-io/VelaTerm vspawn --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/vlinx-io/VelaTerm.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/vspawn .claude/skills/vspawn && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "vspawn" agent skill from https://github.com/vlinx-io/VelaTerm/tree/dev/skills/vspawn into .claude/skills/vspawn/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "vspawn", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/vlinx-io/VelaTerm/tree/dev/skills/vspawnType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add vlinx-io/VelaTerm --skill vspawn -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install vlinx-io/VelaTerm vspawn --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/vlinx-io/VelaTerm.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/vspawn .agents/skills/vspawn && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "vspawn" agent skill from https://github.com/vlinx-io/VelaTerm/tree/dev/skills/vspawn into .agents/skills/vspawn/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "vspawn", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add vlinx-io/VelaTerm --skill vspawn -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install vlinx-io/VelaTerm vspawn --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/vlinx-io/VelaTerm.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/vspawn .cursor/skills/vspawn && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "vspawn" agent skill from https://github.com/vlinx-io/VelaTerm/tree/dev/skills/vspawn into .cursor/skills/vspawn/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "vspawn", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/vlinx-io/VelaTerm.git --path skills/vspawn--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add vlinx-io/VelaTerm --skill vspawn -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install vlinx-io/VelaTerm vspawn --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/vlinx-io/VelaTerm.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/vspawn .gemini/skills/vspawn && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "vspawn" agent skill from https://github.com/vlinx-io/VelaTerm/tree/dev/skills/vspawn into .gemini/skills/vspawn/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "vspawn", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install vlinx-io/VelaTerm vspawnInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add vlinx-io/VelaTerm --skill vspawn -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/vlinx-io/VelaTerm.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/vspawn .github/skills/vspawn && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "vspawn" agent skill from https://github.com/vlinx-io/VelaTerm/tree/dev/skills/vspawn into .github/skills/vspawn/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "vspawn", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add vlinx-io/VelaTerm --skill vspawn -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install vlinx-io/VelaTerm vspawn --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/vlinx-io/VelaTerm.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/vspawn .opencode/skills/vspawn && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "vspawn" agent skill from https://github.com/vlinx-io/VelaTerm/tree/dev/skills/vspawn into .opencode/skills/vspawn/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "vspawn", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
vspawn- GitHub stars
- 259
- Token cost
- ~2.5k tokens
- SKILL.md length
- 1,218 words
- Files
- 3 (incl. references)
- Skills in repo
- 26
- Repo updated
- First seen
- Licence
- MIT
At a glance
Explicitly spawn a standalone child session under the current vlx-term session, passing the task in as its first message (mirrors spawntask).
- Works in 3 steps: Expand the user input into a… → Detect worktree / type options → Run the command
- The user explicitly invokes /vspawn
- SKILL.md covers Planning and execution mode, Step 1: Expand the user input…, Step 2: Detect worktree / type… and Step 3: Run the command, plus 1 more section
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Vspawn is an agent skill from vlinx-io/VelaTerm. Explicitly spawn a standalone child session under the current vlx-term session, passing the task in as its first message (mirrors spawntask). With --plan-execute, create a planner, persistent executors and an optional separate reviewer; --split-tasks enables user-confirmed task decomposition. Runs in the current directory by default, without a git worktree (use vspawn-tree for a worktree). Only use when the user explicitly invokes /vspawn or $vspawn; never auto-trigger. This is a real session run by its own…
Its SKILL.md is about 2.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 3 other files, including reference files (for example `references/plan-execute-legacy.md` and `references/plan-execute.md`).
It sits in Agent Workflows, covering Git worktrees, Subagents and Task breakdown. It works with Git. The repository describes itself as: VelaTerm = Codex + iTerm2, The Best ADE for AI Coding. The licence is MIT.
When your agent uses it
- The user explicitly invokes /vspawn
- Never auto-trigger
Example prompts
- “/vspawn”
Requirements
- Pre-approved tools (allowed-tools): Bash(vspawn:*)
Workflow steps
3 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit 98b5f2f. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves these tools, so the agent can use them without asking each time:
Bash(vspawn:*)
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md (its code samples are bash).
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Vspawn loads about 2.5k tokens when it runs, and up to ~11k if it reads all its reference files. Until then it costs about 168 tokens; SKILL.md has 1,218 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from vlinx-io/VelaTerm at commit 98b5f2f, republished under its MIT licence (© vlinx-io). 1,218 words, ~2,465 tokens.
Download SKILL.mdSave it as
.claude/skills/vspawn/SKILL.md (or your agent's skills folder). This skill also uses 2 other files; get the full folder from GitHub.- name
- vspawn
- description
- Explicitly spawn a standalone child session under the current vlx-term session, passing the task in as its first message (mirrors spawn_task). With --plan-execute, create a planner, persistent executors and an optional separate reviewer; --split-tasks enables user-confirmed task decomposition. **Runs in the current directory by default, without a git worktree** (use vspawn-tree for a worktree). Only use when the user explicitly invokes /vspawn or $vspawn; never auto-trigger. This is a real session run by its own process in the vlx-term left-panel tree — not an in-process sub-agent, and not a background Task. Available only inside vlx-term-hosted sessions.
- allowed-tools
- Bash(vspawn:*)
- argument-hint
- [--plan-execute] [--review|--no-review] [--review-agent <agent>] [--review-model <model>] [--review-effort <level>] [--split-tasks] [--plan-agent <agent>]…
- disable-model-invocation
- true
/vspawn
The user explicitly invoked /vspawn (Claude) or $vspawn (Codex) to request spawning a standalone child
session under the current vlx-term session,
passing the task in as its first message. The new child session runs as a brand-new claude/codex process, and
by default runs in the current directory without a worktree (to open a dedicated git worktree for the child
task, use /vspawn-tree, or add --worktree to this command).
⚠️ This is not an in-process sub-agent, and not a background Task: it is a real, visible, interactive, resumable session tab in the vlx-term left-panel tree, run by its own process, leaving the current session undisturbed.
User input:
$ARGUMENTS
Planning and execution mode
When the user requests this workflow or supplies --plan-execute, keep the
current session as the initiator. Pass the expanded task to vspawn --plan-execute; VelaTerm creates a new
planner, which later dispatches work to its persistent executor. Independent review is disabled by default. Pass --review when the user requests review; it then runs in a separate Review session. Plan only collects context, coordinates and summarizes in both modes. Add --split-tasks when the user requests automatic decomposition into multiple tasks. The new planner then proposes the split for a separate user review; do not split or execute the task in the initiating conversation. Do not perform the planning
or execution here, and do not use in-process subagents.
For a skill invocation that uses both --plan-execute and --split-tasks, automatically add --yes
unless the user explicitly requests the initial configuration dialog. This skips launch configuration and
keeps only the final task-proposal review; the user does not need to supply --yes themselves. Never approve
the proposal on the user's behalf. Without task splitting, keep the ordinary confirmation rules below.
Each role has independent --plan-agent, --plan-model, --plan-effort and --exec-agent, --exec-model,
--exec-effort parameters. Optional Review also has --review-agent, --review-model and --review-effort; --review or --no-review selects whether it exists. Pass only explicit user selections; the backend resolves omitted choices using
the initiating session and supported defaults. With split-task skill invocation, the planner starts with
those settings, and the final review lets the user edit each executor's settings. It does not reconfigure
the already-started planner. Both launch dialogs also offer automatic task splitting. Do not guess a
different model. Only agents supported by the backend chat engine can run this workflow. Agent changes
reset that role's model and effort draft in the dialogs. Ordinary
--model, --effort and agent flags serve as planner defaults when role-specific flags are absent.
Keep the cwd rules below. Map an explicit directory choice to --worktree-mode none (all roles in the
current directory), --worktree-mode shared (one new worktree shared by the planner and all executors), or
--worktree-mode each (one worktree for the planner and a separate worktree for each executor). This option
requires --plan-execute and overrides the legacy --worktree flag. Without an explicit mode, --worktree
means shared; otherwise all workflow sessions use the requested directory. A displayed launch dialog can
change the mode; the final task review retains it. The selected mode is saved for
both single-task and split-task execution and is retained through retries and correction rounds.
Include existing user modifications and the required delivery location in the task. The planner may edit
its plan and delivery-summary documents; only the executor edits implementation files.
The backend supplies all roles with the workflow protocol. Read it when
explaining or diagnosing dispatch, result delivery, correction rounds, cancellation or recovery. In split mode, the user edits and confirms task prompts and execution settings before any executor starts; --yes skips only the initial launch confirmation. Each task has its own workflow ID and round. With review enabled, reports go to the shared reviewer with progress notices to Plan; otherwise the full reports go to Plan. Workflow
messages use vflow actions and vtell --report; they appear with their sender identity in the destination chat. A successful
vspawn invocation records a launch request; it does not prove the user confirmed it or the workflow passed.
Report that the request was submitted and leave this session free. Do not poll after submitting.
Show full SKILL.md (546 more words)Show less
Step 1: Expand the user input into a "self-contained" rich prompt (critical)
The new child session is a brand-new conversation with no memory of this one. Do not forward the user's one-liner verbatim — as if handing the job to someone with zero context, use what is known from the current conversation to expand it into a task description that stands on its own. Write whichever of the following apply — don't pad:
- Goal: what needs to be achieved.
- Relevant files and paths: spell them out (absolute or relative), don't make them guess.
- Established facts / conclusions: key information from this conversation that they wouldn't know (e.g. an already-located root cause, an already-agreed convention).
- Constraints: tech stack, things that must not change, code-style requirements.
- Acceptance criteria: what counts as done.
Don't write "see above / as mentioned / continuing from earlier" — the new session can't see any of this. (If the user's input is already complete, or they explicitly say "keep it short," don't over-expand.)
Step 2: Detect worktree / type options
Detect the following intents from the user input; when detected, turn them into the corresponding command flag, and do not write that instruction into the prompt:
- "want a worktree / open a separate workspace / a separate branch," or a leading
--worktree→ add--worktreeto the command. - Specifying claude / codex, or a leading
--claude/--codex→ add the matching flag. - "no dialog / don't ask me / just start it / start it straight away," or a leading
--yes(-y) → add--yes. This skips the initial launch card, never the split-task proposal review. Split-task skill invocations add it automatically as described above. - Naming a model ("run it on opus / sonnet / gpt-5.5"), or a leading
--model→ add--model <name>. Pass the name through as the user wrote it; model names belong to the agent that will run, and vlx-term turns it into that agent's own model flag. - Naming a reasoning effort ("think harder / low effort / high effort"), or a leading
--effort→ add--effort <level>. Levels are typicallylow,medium,high,xhigh,max; explicit effort selections are rejected if the agent does not expose that setting.
Defaults: no worktree (run in the current directory), follow the current session type, inherit the current session's model and effort, and show the confirmation card when the user has "Confirm before spawn" turned on, except for the split-task skill rule above.
Step 3: Run the command
Choose the command directory before spawning. If the task or conversation identifies one unambiguous project or
repository directory, add --cwd <absolute-path> so the child starts there and a requested worktree is created
from that repository. Otherwise omit it and let vspawn use the command's current directory. Do not guess between
multiple plausible repositories.
Pass the prompt you expanded in Step 1 as a single argument (escaping any quotes inside it correctly), and run
vspawn from PATH:
vspawn [--worktree] [--cwd <absolute-path>] [--yes] [--claude|--codex] [--model <name>] [--effort <level>] "<expanded self-contained prompt>"For a split-task skill invocation, use this form, adding only the user's explicit role and directory choices:
vspawn --plan-execute --split-tasks --yes [--worktree-mode <none|shared|each>] [--cwd <absolute-path>] [--plan-agent <agent>] [--exec-agent <agent>] "<expanded self-contained prompt>"After it succeeds, give a one-line summary that the launch request was submitted, including the task and whether it uses planning and execution mode. A submitted request may still be awaiting confirmation.
Notes
- Must run inside a vlx-term-hosted session: it relies on the injected
VLX_*environment variables and thevspawnon PATH; if missing it reports "not inside a vlx-term session" and exits.
© vlinx-io, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
SKILL.md and 2 other files (references) in skills/vspawn of vlinx-io/VelaTerm.
- SKILL.md
- references/plan-execute-legacy.md
- references/plan-execute.md
Open the folder on GitHubat commit 98b5f2f
Compare with similar skills
Vspawn next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Vspawn this skillvlinx-io/VelaTerm | 259 | — | ~2.5k | Automated safety check: Pass | MIT | |
| Agtx Task Sweepfynnfluegge/agtx | 1.7k | — | ~1.7k | Automated safety check: Pass | Apache-2.0 | |
| PmApra-Labs/apra-fleet | 101 | — | ~5.4k | Automated safety check: Pass | Custom licence | |
| Vibe Cherry Pick Integrationash1794/vibe-engineering | 162 | — | ~573 | Automated safety check: Pass | MIT | |
| Qwen Code Team CoordinatorQwenLM/qwen-code | 28k | — | ~994 | Automated safety check: Pass | Apache-2.0 | |
| Agenthubalirezarezvani/claude-skills | 28k | — | ~2k | Automated safety check: Pass | MIT |
Similar skills
Pm
Apra-Labs/apra-fleet
Project Manager skill. An agent skill from Apra-Labs/apra-fleet.
Vibe Cherry Pick Integration
ash1794/vibe-engineering
Safely integrates commits from parallel agent branches using sequential cherry-pick.
Qwen Code Team Coordinator
QwenLM/qwen-code
Leads a small team of Qwen Code teammates: read-only investigators work in parallel, then one writer pinned to a git worktree makes the changes.
Agenthub
alirezarezvani/claude-skills
Multi-agent collaboration plugin that spawns N parallel subagents competing on the same task via git worktree isolation.
CCPM Project Management
automazeio/ccpm
Runs a spec-driven workflow from PRD to epic to GitHub issues to parallel agents, with status, standup and blocked-work reports from bundled scripts.
More from vlinx-io/VelaTerm
All 26 skills in this repoAssess Patch Risk
vlinx-io/VelaTerm
Assess an immutable patch artifact's program impact, regression risk, and auto-merge eligibility.
Deep Security Scan
vlinx-io/VelaTerm
A skill your agent uses when the user asks for a deep, exhaustive, multi-pass, or variance-reducing repository-wide or scoped-path Codex Security scan.
Define Security Policy
vlinx-io/VelaTerm
Define, review, or update SECURITY.md guidance for a repository or component.
Track Findings
vlinx-io/VelaTerm
Track validated Codex Security findings in Linear, Jira, GitHub issues, or draft GitHub security advisories.
Verify Fix
vlinx-io/VelaTerm
Use only when the user explicitly requests verification that a security fix remediates a reported vulnerability.
Vopen
vlinx-io/VelaTerm
Open a file or URL in the vlx-term center pane (mirrors the vopen command).
Related
Works with
Categories
Questions about Vspawn
What does Vspawn do?
Explicitly spawn a standalone child session under the current vlx-term session, passing the task in as its first message (mirrors spawntask). Vspawn is an agent skill from vlinx-io/VelaTerm. Explicitly spawn a standalone child session under the current vlx-term session, passing the task in as its first message (mirrors spawntask).
When should I use Vspawn?
Vspawn fits situations like: the user explicitly invokes /vspawn; never auto-trigger.
How do I install Vspawn in Claude Code?
Run `npx skills add vlinx-io/VelaTerm --skill vspawn -a claude-code`. Or copy the skill folder (skills/vspawn in vlinx-io/VelaTerm) into .claude/skills/vspawn in your project. Claude Code loads it when a task matches its description.
How do I install Vspawn in Codex?
Run `npx skills add vlinx-io/VelaTerm --skill vspawn -a codex`. Or copy the skill folder (skills/vspawn in vlinx-io/VelaTerm) into .agents/skills/vspawn in your project. Codex loads it when a task matches its description.
Can I use Vspawn in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add vlinx-io/VelaTerm --skill vspawn -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/vspawn, .gemini/skills/vspawn, .github/skills/vspawn and .opencode/skills/vspawn in your project.
What does Vspawn need to run?
SKILL.md names no scripts, command-line tools or credentials: Vspawn is instructions for the agent only. Its frontmatter pre-approves these tools: Bash(vspawn:*).
Does Vspawn access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Vspawn safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Vspawn use?
Vspawn is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Vspawn use?
About 2.5k tokens (SKILL.md is roughly 9.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 8.1k tokens, read only when the agent opens those files.
What are the alternatives to Vspawn?
Skills that share tags, products or a category with Vspawn: Agtx Task Sweep (fynnfluegge/agtx, 1.7k stars), Pm (Apra-Labs/apra-fleet, 101 stars), Vibe Cherry Pick Integration (ash1794/vibe-engineering, 162 stars) and Qwen Code Team Coordinator (QwenLM/qwen-code, 28k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Vspawn?
vlinx-io (a GitHub user) maintains it in vlinx-io/VelaTerm, which has 259 GitHub stars. The repository holds 26 skills in this directory. The repository was last updated on October 7, 2026.
Source: vlinx-io/VelaTerm on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.