Skip to content
Agentmart
Submit a tool

agent-health

External health monitor for Loom (Rising Leviathan).

Skill by arc0btc

IndexedMIT★ 1
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/arc0btc-agent-health

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

External health monitor for Loom (Rising Leviathan). Analyzes pre-gathered cycle metrics, task failures, git activity, and gate state to classify agent health and send email alerts.

Use cases: productivity, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found
Runs as
Node.js package

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/agent-health

What's inside

Skill name
agent-health
Description
External health monitor for Loom (Rising Leviathan). Analyzes pre-gathered cycle metrics, task failures, git activity, and gate state to classify agent health and send email alerts.
License
MIT
Instructions
139 lines in SKILL.md · view on GitHub

Before installing

  • Ships executable scripts

Files (3)

  • SKILL.md
  • loom.json · agents/
  • sensor.ts

SKILL.md (first lines)

# Agent Health Monitor

## Purpose

This skill is loaded by Haiku tasks created by the `agent-health` sensor running on Arc (Trustless Indra). Arc SSHes into Loom every 2 hours, gathers all relevant data, and bakes it into the task description as a structured data block.

**Your only job as Haiku:** read the pre-gathered data, classify health status (GREEN / YELLOW / RED), and send an email if anything needs attention. You do not need to SSH anywhere or query any databases -- everything is already in the task description.

## Data Block Format

The task description will contain a block structured like this:

```
=== AGENT HEALTH DATA: Loom (Rising Leviathan) ===
Checked: {ISO timestamp}
Period: last {N} hours

## Cycle Metrics
cycles_checked: N
total_tokens_in: N
total_cost_usd: N.NN
avg_tokens_per_cycle: N
avg_cost_per_cycle: N.NN
max_tokens_in_single_cycle: N
max_cost_single_cycle: N.NN
spike_cycles: [{started_at, tokens_in, cost_usd, model, duration_ms, task_subject}]

## Task Failures
failed_sources: [{source, count, last_subject}]
retry_storms: [{source, count, subjects}]

## Git Activity
commits_since_last_check: N
watched_path_commits: [{hash, date, message, files_changed}]

## Gate State
dispatch_gate: {status, consecutive_failures, stopped_at, stop_reason}
watchdog: {last_ran, last_alert_at, last_result}

## Thresholds
tokens_in_per_cycle: 1000000
cost_per_cycle_usd: 3.00
daily_cost_usd: 100
failed_source_repeat_count: 5
pending_task_age_hours: 4
=== END DATA ===
```

## Analysis Instructions

Work through these steps in order:

### Step 1: Compare Each Metric Against Its Threshold

Check each of the following, noting which are within bounds and which exceed thresholds:

| Metric | Threshold | Signal |
|--------|-----------|--------|
| max_tokens_in_single_cycle | 1,000,000 | RED if exceeded |
| max_cost_single_cycle | $3.00 | RED if exceeded |
| total_cost_usd (for period) | $100/day equivalent | YELLOW if on pace to exceed |
| failed_source repeat count | 5 occurrences | RED if any source exceeds this |
| watched_path git commits | any non-trivial changes | YELLOW if behavior-critical code changed |
| pending task age | 4 hours | YELLOW if any task pending longer |
| dispatch_gate status | "open" | RED if stopped or closed |

### Step 2: Apply Known Noise Filters

Skip the following when counting failures (these sensors create/fail tasks as normal operation):
- `sensor:treasury-health`
- `sensor:service-health`
- `sensor:service-health:stale-lock`

Do not flag these sources as retry storms even if their counts are high.

### Step 3: Classify Overall Status

- **GREEN**: All metrics within thresholds, no watched-path commits, gate open
- **YELLOW**: At least one metric in warning range, OR non-trivial watched-path commits, OR any pending task past age threshold
- **RED**: Any threshold critically exceeded, OR dispatch gate stopped, OR retry storm detected from a non-noise source

If multiple signals exist, use the worst classification.

### Step 4: Decide Whether to Send Email

- **GREEN**: Do not send email. Mark the task complete with a one-line "all clear" note.
- **YELLOW**: Send email to whoabuddy@gmail.com with subject `[Loom Health] YELLOW - {brief reason}`.
- **RED**: Send email to whoabuddy@gmail.com with subject `[Loom Health] RED - {brief reason}`.

### Step 5: Compose Email Body

For YELLOW emails, include:
- Overall status classification
- Which metrics triggered the warning
- Relevant data points (not the full dump, just what's relevant)
- Suggested investigation steps

For RED emails, include:
- Overall status classification (urgent)
- Which thresholds were exceeded and by how much
- Full data for the flagged metrics (paste the relevant sections from the data block)
- Immediate action items (e.g., "check if Loom is running a token spiral", "inspect recent git commits to src/")

### Step 6: Send Email Using arc-email-sync

Use the `arc-email-sync` skill to send the email:
- To: whoabuddy@gmail.com
- From: arc@arc0.me (default)
- Subject: as defined above
- Body: as composed above

## Email Quick Reference

| Status | Action | Subject Pattern |
|--------|--------|-----------------|
| GREEN | No email, note "all clear" | (none) |
| YELLOW | Send warning email | `[Loom Health] YELLOW - {reason}` |
| RED | Send urgent email | `[Loom Health] RED - {reason}` |

From arc0btc/arc-starter, MIT.

More from arc0btc/arc-starter

Signals

Stars
1
Forks
4
Open issues
3
Last commit (this folder)
Jul 14, 2026 (3 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from arc0btc

Cite this page

agent-health skill. Agent Mart. https://agentmart.dev/skills/arc0btc-agent-health (last updated 2026-10-24)