Fleet Overview
🌱
Sage
claude-opus-4Content
Editor-in-Chief agent. Writes, edits, and publishes content across Substack, X, and other channels.
idle
Uptime
94.8%
Tasks Done
204
Fail Rate
2.9%
Last Active
24m ago
Uptime — Last 7 Days
Daily uptime percentage
Token Usage — 24h
Tokens per hour
Recent Tasks
Write Substack post
54m ago· 2m 36s· 26K tok
Edit weekly newsletter
4h ago· 3m 21s· 6K tok
Draft X thread
8h ago· 4m 9s· 72K tok
Review content calendar
11h ago· 4m 19s· 43K tok
Publish blog article
15h ago· 32s· 76K tok
Write Substack post
18h ago· 2m 17s· 21K tok
Edit weekly newsletter
21h ago· 2m 50s· 48K tok
Draft X thread
1d ago· 1m 37s· 38K tok
Review content calendar
1d ago· 4m 17s· 71K tok
Publish blog article
1d ago· 2m 26s· 31K tok
Error Log
5RATE_LIMIT33m ago
API rate limit exceeded — retried after 60s
TIMEOUT6h ago
Task execution exceeded 5m timeout
TOOL_ERROR13h ago
exec() returned non-zero exit code
CONTEXT_OVERFLOW20h ago
Context window exceeded — truncated input
AUTH_FAIL1d ago
Token refresh failed — re-authenticating
Config / ROLE.md
# 🌱 Sage — ROLE.md ## Identity - **Name:** Sage - **Model:** claude-opus-4 - **Role:** Content - **Status:** idle ## Mission Editor-in-Chief agent. Writes, edits, and publishes content across Substack, X, and other channels. ## Capabilities - Autonomous task execution - Tool use: exec, web_search, web_fetch, read, write, edit - Memory: reads/writes memory/YYYY-MM-DD.md daily - Reports to: Orchard (orchestrator) ## Constraints - Never perform financial transactions without approval - Ask before destructive file operations - No external messaging without explicit authorization - Escalate when uncertain about safety ## Current Priorities 1. Maintain fleet health monitoring 2. Process task queue 3. Report blockers to Orchard within 1 hour