auto-task

auto-task is a command for coding agents from rube-de/cc-skills. It costs 42 tokens per session (2,703 once invoked), scanned C, original, MIT.

An autonomous workflow command that coordinates separate agents for planning, implementation, testing, review, and reporting. An agent is a software worker assigned a specific role in the development process.

In plain words
What is it for?
Use it to run a complete software task involving architecture planning, product planning, coding, code tests, quality checks, review, and a final report.
Why use it?
It organizes a multi-stage coding task so the work can move from a plan to development and checks without an approval pause between those phases.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/rube-de/cc-skills/auto-task
Clone the repo
git clone --depth 1 https://github.com/rube-de/cc-skills

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for auto-task

README.md
[![agentmods](https://agentmods.dev/badge/commands/rube-de/cc-skills/auto-task.svg)](https://agentmods.dev/commands/rube-de/cc-skills/auto-task)
Your own site
<a href="https://agentmods.dev/commands/rube-de/cc-skills/auto-task"><img src="https://agentmods.dev/badge/commands/rube-de/cc-skills/auto-task.svg" alt="Measured on agentmods" height="20"></a>
Per session 42 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,703 The whole file, excluding the scripts and references it only reads on demand.
Security scan C 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00042 $0.02703
Opus 5 $0.00021 $0.01352
Sonnet 5 $0.00008 $0.00541
Haiku 4.5 $0.00004 $0.00270

Measured 4d ago against content hash 432972202892, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade C, and why

auto-task scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

10. Clean up branch state: `[ -n "$BRANCH_SLUG" ] && rm -rf ".dev/cdt/$BRANCH_SLUG"`
plugins/cdt/commands/auto-task.md · 201 lines

How it starts

The opening of the file, as written. The whole thing — 201 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ROLE: Coordinator only. You MUST NOT use Edit, Write, or NotebookEdit tools on source code, test files, or project docs. All implementation, testing, review, and documentation MUST be delegated to teammates via SendMessage. If you find yourself about to edit a file, STOP and delegate to the appropriate teammate instead. You verify plan/report artifacts written by teammates. Narrow exception: One-shot Bash file-appends for plugin-install side effects (specifically the discovery-hint install performed in Wrap Up step 8) are explicitly permitted. This exception does NOT extend to Edit/Write/NotebookEdit on any file, nor to broader Bash file edits on source, test, or doc content.

/auto-task — Autonomous Workflow

Target: $ARGUMENTS

Two-phase orchestration like /cdt:full-task, but without a user approval gate. Plan team runs, cleans up, then dev team starts immediately.

Phase 1: plan-team                    Phase 2: dev-team
┌───────────────────────────┐         ┌───────────────────────────┐
│  Lead (You)               │         │  Lead (You)               │
│  ├── architect       [tm] │ plan.md │  ├── developer   [tm]     │
│  ├── product-manager [tm] │──────→  │  ├── code-tester [tm]     │
│  └── researcher      [sa] │         │  ├── qa-tester   [tm]     │
└───────────────────────────┘         │  ├── reviewer    [tm]     │
                                      │  └── researcher  [sa]     │
                                      └───────────────────────────┘

[tm] = teammate (Agent Team) [sa] = subagent

The plan-team / dev-team labels above are role names; the actual team names created by the workflow are scoped per run as plan-<branch-slug>-<timestamp> and dev-<branch-slug>-<timestamp> to avoid collisions on the global ~/.claude/teams/ namespace, where branch-slug is the current branch name with / replaced by -.

Step 0: Git Check

  1. Run git fetch origin
  2. Ensure you are on main or master — if not, run git checkout main (or master, whichever exists)
  3. Suggest a branch name based on the task (e.g. feat/rate-limiting)
  4. Run git checkout -b <branch> origin/<default-branch> to create the feature branch from the latest remote default branch
  5. Run git pull to ensure up-to-date

Read the full file on GitHub · 201 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago First seen · 201 lines · 42 tokens per session scan C 432972202892

Subscribe to this mod's changes

auto-task is a command published in the GitHub repository rube-de/cc-skills (10 stars, last pushed today), licensed MIT. It adds 42 tokens to every session and 2,703 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.