- Home
- Skills
- Workflow Automation
- safe autopilot
Works with the AI Agents you already use
safe autopilot
Build and audit safe, scheduled automations with dry-run modes, undo journals, and failure alerts.
$2
safe autopilot
Example session with this skill installed
I need a script to move files older than 30 days from /logs to a /trash folder. It needs to run every morning at 4 AM on my Linux server.
- Read your context and instructions
- Compiled the safe autopilot
I have scaffolded a new files automation project and classified this as an L1 risk (Bulk file move). The logic is implemented in run.py using ctx.trash() to ensure nothing is permanently deleted, and I have generated a systemd timer for your 4 AM schedule.
Safe File Cleanup Automation
# run.py - Task implementation
def task(ctx, cfg, state):
log_dir = cfg.get("log_dir", "/logs")
retention_days = cfg.get("retention_days", 30)
# Calculate cutoff
import time
cutoff = time.time() - (retention_days * 86400)
for item in ctx.list_dir(log_dir):
if item.mtime < cutoff:
# ctx.trash moves to a local .trash folder and journals the action
ctx.trash(item.path)
ctx.stats["cleaned_files"] += 1
ctx.notify(f"Trashed expired log: {item.name}")
Generated Schedule (systemd)
# /etc/systemd/system/log-cleanup.timer
[Unit]
Description=Daily cleanup of logs older than 30 days
[Timer]
OnCalendar=*-*-* 04:00:00
Persistent=true
[Install]
WantedBy=timers.target
Next steps
- Run
python3 run.pyin the project folder to perform a dry run and verify which files will be moved. - Review the
config.example.jsonto ensure the/logspath is correct for your environment. - Once verified, run
python3 run.py --applyto perform the first supervised cleanup.
Connects securely to your tools. The creator never sees your data.
About this skill
Automations fail in ways that cost time and data. Scripts often overwrite files, hang indefinitely, or fail silently without anyone noticing. This skill provides a rigorous framework for building and auditing scheduled tasks that are safe to leave running unattended.
What it does
- Project scaffolding generates complete Python projects with dry-run modes, undo journals, and file trashing instead of deletion.
- Safety auditing scans existing scripts, crontabs, and GitHub workflows for dangerous patterns like hard-coded secrets or missing timeouts.
- Idempotent runtime provides a standard kit for locking processes, rotating logs, and handling retries with backoff.
- Schedule generation converts plain-English intervals into valid cron, systemd, launchd, or Windows Task Scheduler configurations.
- Monitoring setup integrates failure alerts via webhooks and heartbeats to notify you when a task stops running.
How it works
- Classify risk by identifying the job trigger, frequency, and potential cost of failure using a standardized risk ladder.
- Scaffold or audit by generating a new
autokitproject or running the auditor on your existing shell, Python, or JS scripts. - Implement logic using safe context methods for file operations that automatically support dry-runs and undo functionality.
- Validate and schedule by running the built-in test suite and generating a target-specific schedule file.
Frameworks & tools
The generated projects require Python 3.8 or higher with no external package dependencies. It supports scheduling for Linux (cron, systemd), macOS (launchd), Windows (Task Scheduler), and GitHub Actions.
Why this beats prompting it yourself
Generic LLM prompts often produce scripts that lack error handling, state management, or safety checks. This skill enforces a "dry-run by default" architecture and provides an undo journal that a simple script cannot replicate.
Use cases
- File management such as sorting downloads, cleaning up expired backups, or syncing folders without risk of data loss.
- System monitoring like health-checking URLs or services and alerting only when status changes.
- Cloud workflows including GitHub Actions that need time-zone-aware scheduling and failure notifications.
- Script hardening for legacy cron jobs that currently lack logging, locking, or error reporting.
Known limitations
The undo functionality only covers file operations performed through the provided context methods. It cannot reverse external API calls, sent emails, or database writes.
Undo functionality only covers file operations via the ctx tool. Requires Python 3.8+. Does not reverse external API calls or database writes.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 10 days ago
Requires
Permissions
Allowed Hosts