Skip to content
Get packs

Autonomous Agent Harness

Turns Claude Code into a scheduled, memory-keeping agent that runs recurring tasks

Source Download skill
Repository
affaan-m/ECC
Stars
295,127
Licence
MIT
Files
1 file

What it does

This skill teaches Claude Code to act as a long-running assistant: it describes how to use Claude Code's built-in memory, scheduled tasks (cron-style jobs), remote agent dispatch, and MCP memory servers so the agent can keep context across sessions and repeat work on a timetable. It is written in Japanese and is a set of guidelines and example configurations rather than a packaged program. It also states that automation should only be set up when the user explicitly asks, with a defined scope, and that dry-run planning comes first.

Great for

  • Developers who want Claude Code to run a recurring check, such as a daily pull request review
  • Users who want to replicate features of standalone agent frameworks like AutoGPT inside Claude Code
  • Teams experimenting with scheduled, unattended agent workflows

Skip it if

  • People who want a simple one-off task; the setup adds scheduling and memory that persist and can act without supervision
  • Anyone handling sensitive accounts or private data without reviewing what the scheduled prompts can access, since the skill covers credentials and private workspaces only briefly
  • Users who need computer control or posting to external services without careful oversight, because the skill itself mentions these as powerful actions that need explicit approval

Try asking

  • “Set up a weekday 9am job that reviews open pull requests and posts a summary to memory”
  • “Create a daily task that checks a repository's CI status and flags problems”
  • “Plan an autonomous monitoring loop with a dry-run first before enabling it”
  • “Help me build a personal assistant that remembers my project preferences across sessions”
Step-by-step guide: how to use Autonomous Agent HarnessInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    autonomous-agent-harness.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip autonomous-agent-harness.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip autonomous-agent-harness.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Start with a dry-run plan and a local queue file before enabling any scheduled job, as the skill recommends
  • Scope each scheduled task narrowly (specific repo, specific actions) and state explicitly what it may and may not post or change
  • Review the memory store periodically, since it persists across sessions and may hold sensitive details

What's inside

  • SKILL.md

Redistributed unmodified from affaan-m/ECC under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.