autonomous-longtask

autonomous-longtask is a skill for Claude Code, Codex from T-0-co/t-0-spec-kit-ralph. It costs 38 tokens per session (2,244 once invoked), scanned A, original, MIT.

A German-language guide for handling large software-development tasks across multiple sessions. It covers tested increments, checkpoints, sub-agents, parallel work, and progress handoffs.

In plain words
What is it for?
It is for complex features, multi-step refactorings, work spanning several sessions, and tasks that take more than 30 minutes.
Why use it?
It helps reduce unfinished work, lost context, and untested changes during long development tasks.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: mentions subagents; names the TodoWrite tool; mentions Claude Code.

Needs its repository: it reads a path above its own folder, which exists only inside the repository. The line is git worktree add ../feature-a feature/a.

Good fit It is for complex features, multi-step refactorings, work spanning several sessions, and tasks that take more than 30 minutes.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.

Clone the repo
git clone --depth 1 https://github.com/T-0-co/t-0-spec-kit-ralph
agentmods
npx agentmods add skills/t-0-co/t-0-spec-kit-ralph/autonomous-longtask

Made for: Claude Code, Codex.

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 autonomous-longtask

README.md
[![agentmods](https://agentmods.dev/badge/skills/t-0-co/t-0-spec-kit-ralph/autonomous-longtask/github.svg)](https://agentmods.dev/skills/t-0-co/t-0-spec-kit-ralph/autonomous-longtask)
Your own site
<a href="https://agentmods.dev/skills/t-0-co/t-0-spec-kit-ralph/autonomous-longtask"><img src="https://agentmods.dev/badge/skills/t-0-co/t-0-spec-kit-ralph/autonomous-longtask/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for autonomous-longtask

Your own site · 80×15
<a href="https://agentmods.dev/skills/t-0-co/t-0-spec-kit-ralph/autonomous-longtask"><img src="https://agentmods.dev/badge/skills/t-0-co/t-0-spec-kit-ralph/autonomous-longtask.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 38 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,244 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00038 $0.02244
Opus 5 $0.00019 $0.01122
Sonnet 5 $0.00008 $0.00449
Haiku 4.5 $0.00004 $0.00224

Measured 10d ago against content hash e432ce96f1db, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

autonomous-longtask scanned grade A with 0 findings 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 10d 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.agents/skills/autonomous-longtask/SKILL.md · 379 lines

How it starts

The opening of the file, as written. The whole thing — 379 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Autonomous Long-Task Development

Dieses Skill optimiert Claude Code für lange, autonome Entwicklungsaufgaben - von mehrstündigen Feature-Implementierungen bis zu Multi-Session-Refactorings.

Wann diesen Skill nutzen

  • Komplexe Features: Implementierungen die mehrere Dateien/Services betreffen
  • Multi-Step Workflows: Tasks mit 5+ abhängigen Schritten
  • Long-Running Tasks: Aufgaben die >30 Min dauern
  • Multi-Session Tasks: Arbeit die über Context-Grenzen hinausgeht

Core Principles

1. Loop Closing (Test-Driven)

Jede Code-Änderung muss einen Feedback-Loop schließen:

1. Test schreiben (was soll passieren)
2. Code implementieren
3. Test ausführen → Loop geschlossen
4. Refactor falls nötig
5. Test erneut → Sicherheit

Anti-Pattern (Open Loop):

"Implementiere Feature X"
→ Code geschrieben, keine Verifikation
→ Bugs später entdeckt

Best Practice (Closed Loop):

"Implementiere Feature X mit E2E Test.
Führe den Test nach Implementation aus."
→ Sofortige Verifikation

2. Incremental Progress mit Checkpoints

Niemals alles auf einmal:

❌ Versuchen das komplette Feature zu "one-shotten"
   → Context läuft aus mitten in der Implementierung
   → Nächste Session erbt Chaos

✓ Kleine, getestete Inkremente
   → Jedes Inkrement funktioniert standalone
   → Klare Übergabe zwischen Sessions

Checkpoint-Pattern:

  • Nutze /rewind oder Esc Esc für Rollbacks
  • Committe nach jedem funktionierenden Inkrement
  • Dokumentiere State in STATUS.md oder claude-progress.txt

3. Context-Window Management

Claude Opus 4 kann 200K Tokens, aber:

  • Sub-Agents haben eigene Context Windows (isoliert)
  • Lange Sessions fragmentieren den Context
  • claude-progress.txt für Session-Handoffs nutzen

Sub-Agents für Parallelisierung

Was sind Sub-Agents?

Lightweight Claude-Instanzen mit eigenem Context Window. Nur relevante Ergebnisse kommen zurück zum Orchestrator.

Verfügbare Agent-Typen

Read the full file on GitHub · 379 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. 10d ago First seen · 379 lines · 38 tokens per session scan A e432ce96f1db

Subscribe to this mod's changes

autonomous-longtask is a skill published in the GitHub repository T-0-co/t-0-spec-kit-ralph (10 stars, last pushed 2mo ago), licensed MIT. It adds 38 tokens to every session and 2,244 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.