Troubleshooting¶
Quick Diagnostics¶
This checks engine, agents, Telegram token, LLM connection, and reports any issues.
Common Issues¶
Bot doesn't start¶
Check:
Fix:
# Reconfigure
simplecontext-bot setup
# Or start fresh
rm -rf ~/.simplecontext-bot
simplecontext-bot setup
LLM errors¶
❌ Invalid API key
⚠️ Rate limit reached
You've hit the API rate limit. Wait a moment, then try again. Consider switching to a different model or provider.
❌ Connection error
Check your internet connection. For Ollama: make sure Ollama is running (ollama serve).
Agents not loading¶
Symptoms: Bot responds with general agent for everything, or /agents shows an empty list.
# Check if agents are installed
simplecontext-bot status
# Reinstall agents
simplecontext-bot update --agents-only
Plugin not loading¶
Symptoms: Plugin not in /plugins, plugin commands missing from /help.
Check the plugin file:
python3 -c "
import sys
sys.path.insert(0, '/path/to/simplecontext')
from vector_search_plugin import VectorSearchPlugin
print(VectorSearchPlugin)
"
Check logs:
Common causes: - Plugin file has a syntax error - Plugin class doesn't inherit BasePlugin - Plugin class doesn't have a name attribute - Import error in plugin file
Memory not persisting¶
Symptoms: Bot forgets everything on restart.
Check that bot.db exists:
If missing, the storage path may be wrong:
Agents hot-reload not working¶
Hot-reload watches the agents/ folder for file changes. Make sure:
hot_reload: truein config (check~/.simplecontext-bot/config.json)- You're editing files in the right folder:
~/.simplecontext-bot/agents/
/semantic command not available¶
The /semantic command requires the vector-search plugin:
simplecontext-bot plugins install vector-search
simplecontext-bot start # restart to register the command
Reset and Start Fresh¶
This removes all config, memory, agents, and plugins. Run setup again from scratch.
Debug Logging¶
Debug mode logs: - Agent routing decisions - Context retrieval pipeline (candidates, scores, selection) - Plugin hook calls - LLM request/response