Autonomous Agent Harness
Install this when you want Claude Code to run on a schedule, remember context across sessions, and queue work like a personal AutoGPT without a separate framework repo.
Overview
Autonomous Agent Harness is an agent skill most often used in Build (also Operate infra, Grow lifecycle) that configures Claude Code for persistent memory, scheduled tasks, and self-directing agent loops via native crons
Install
npx skills add https://github.com/affaan-m/everything-claude-code --skill autonomous-agent-harnessWhat is this skill?
- Maps native Claude Code crons, dispatch, MCP tools, and memory into one autonomous loop
- Explicit consent and safety boundaries before schedules, computer use, or external posts
- Dry-run plans and local queue files recommended before recurring automation
- Replaces standalone Hermes/AutoGPT-style setups by staying inside ECC primitives
- Computer-use and monitoring patterns for always-on personal assistant workflows
Adoption & trust: 3.4k installs on skills.sh; 210k GitHub stars; 1/3 security scanners passed (skills.sh audits).
What problem does it solve?
You want an agent that keeps working between sessions, but standalone autonomous frameworks feel heavy and misaligned with how Claude Code already runs tools.
Who is it for?
Power users who explicitly want scheduled or monitoring-style Claude Code operation with clear safety scopes and ECC-native features.
Skip if: Casual one-off coding tasks, unapproved scraping or posting, or teams that cannot define workspace boundaries and credential handling.
When should I use this skill?
The user wants continuous autonomous operation, scheduled tasks, monitoring language like run every day, or a self-directing agent loop inside Claude Code.
What do I get? / Deliverables
You get a consent-scoped harness plan—memory, queues, schedules, and MCP hooks—so approved automations run continuously without exporting private data into reusable artifacts.
- Harness configuration plan with crons, queues, and memory layout
- Dry-run queue files before live schedules
- Documented consent boundaries for computer use and external actions
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Persistent harnesses, crons, dispatch, and MCP wiring are built artifacts—solo builders assemble them while shaping their agent stack. Agent-tooling is the shelf for turning Claude Code into a self-directing system using native crons, memory, and task queues.
Where it fits
Configure crons, dispatch, and MCP tools into a single autonomous loop for your repo.
Schedule nightly log or error summaries the user approved for their own services.
Queue recurring content or support triage steps after explicit scope sign-off.
Dry-run automation plans in local queue files before enabling recurring production actions.
How it compares
ECC-native orchestration skill—not a drop-in Hermes server or unmanaged AutoGPT deployment.
Common Questions / FAQ
Who is autonomous-agent-harness for?
Solo builders and indie hackers who treat Claude Code as their primary agent runtime and want durable schedules, queues, and memory without another framework repo.
When should I use autonomous-agent-harness?
Use it in Build when wiring crons and MCP for your assistant, in Operate when setting up regular health checks, and in Grow when automating repeatable lifecycle follow-ups the user approved.
Is autonomous-agent-harness safe to install?
The skill mandates explicit user consent for sensitive actions; review the Security Audits panel on this page and never enable computer use or external writes without scoped approval.
SKILL.md
READMESKILL.md - Autonomous Agent Harness
# Autonomous Agent Harness Turn Claude Code into a persistent, self-directing agent system using only native features and MCP servers. ## Consent and Safety Boundaries Autonomous operation must be explicitly requested and scoped by the user. Do not create schedules, dispatch remote agents, write persistent memory, use computer control, post externally, modify third-party resources, or act on private communications unless the user has approved that capability and the target workspace for the current setup. Prefer dry-run plans and local queue files before enabling recurring or event-driven actions. Keep credentials, private workspace exports, personal datasets, and account-specific automations out of reusable ECC artifacts. ## When to Activate - User wants an agent that runs continuously or on a schedule - Setting up automated workflows that trigger periodically - Building a personal AI assistant that remembers context across sessions - User says "run this every day", "check on this regularly", "keep monitoring" - Wants to replicate functionality from Hermes, AutoGPT, or similar autonomous agent frameworks - Needs computer use combined with scheduled execution ## Architecture ``` ┌──────────────────────────────────────────────────────────────┐ │ Claude Code Runtime │ │ │ │ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌─────────────┐ │ │ │ Crons │ │ Dispatch │ │ Memory │ │ Computer │ │ │ │ Schedule │ │ Remote │ │ Store │ │ Use │ │ │ │ Tasks │ │ Agents │ │ │ │ │ │ │ └────┬─────┘ └────┬─────┘ └────┬─────┘ └──────┬──────┘ │ │ │ │ │ │ │ │ ▼ ▼ ▼ ▼ │ │ ┌──────────────────────────────────────────────────────┐ │ │ │ ECC Skill + Agent Layer │ │ │ │ │ │ │ │ skills/ agents/ commands/ hooks/ │ │ │ └──────────────────────────────────────────────────────┘ │ │ │ │ │ │ │ │ ▼ ▼ ▼ ▼ │ │ ┌──────────────────────────────────────────────────────┐ │ │ │ MCP Server Layer │ │ │ │ │ │ │ │ memory github exa supabase browser-use │ │ │ └──────────────────────────────────────────────────────┘ │ └──────────────────────────────────────────────────────────────┘ ``` ## Core Components ### 1. Persistent Memory Use Claude Code's built-in memory system enhanced with MCP memory server for structured data. **Built-in memory** (`~/.claude/projects/*/memory/`): - User preferences, feedback, project context - Stored as markdown files with frontmatter - Automatically loaded at session start **MCP memory server** (structured knowledge graph): - Entities, relations, observations - Queryable graph structure - Cross-session persistence **Memory patterns:** ``` # Short-term: current session context Use TodoWrite for in-session task tracking # Medium-term: project memory files Write to ~/.claude/projects/*/memory/ for cross-session recall # Long-term: MCP knowledge graph Use mcp__memory__create_entities for permanent structured data Use mcp__memory__create_relations for relationship mapping Use mcp__memory__add_observati