Clawdbot Logs Litiao

Analyze Clawdbot logs and diagnostics. Use when the user asks about bot performance, response times, errors, session stats, token usage, API costs, or wants...

Security Scan

The skill's files and instructions match its stated purpose (log and session diagnostics); it reads local journals and session files but does not contact external endpoints or request unrelated credentials.

Purpose & Capability

Name/description align with requested actions: the SKILL.md and included scripts only read clawdbot journals, user systemd service status, and files under ~/.clawdbot to report response times, errors, token usage and costs. One minor mismatch: SKILL.md uses common system tools (journalctl, systemctl, jq, grep, tail) but the skill declares no required binaries—this is an omission but not a functional incoherence.

Instruction Scope

Instructions and scripts explicitly read sensitive local data (session JSON/JSONL containing full conversation history, token/cost info) and the user journal for the clawdbot service. All referenced paths and commands are within the expected scope for diagnostics; there are no steps that transmit data externally. Users should be aware that running this skill exposes local conversation contents to whoever or whatever is invoking the skill.

Install Mechanism

Instruction-only skill with two small shell scripts included. No install spec, no downloads, and nothing is written to disk beyond running the provided scripts. This is low-risk from an installation perspective.

Credentials

The skill requests no environment variables or credentials, which is appropriate. However it reads files in $HOME and queries the user journal/systemd user service; access to these filesystem and journal entries grants visibility into sensitive conversation content and token/cost data—this is expected for a diagnostics tool but worth highlighting.

Persistence & Privilege

always is false and there is no install-time privilege escalation or modification of other skills or system-wide config. The skill does not request persistent platform privileges beyond being invoked.

Assessment

This skill appears to do exactly what it claims: read Clawdbot journals and session files and summarize performance and cost data. Before installing or running it, remember: (1) it reads local session files (~/.clawdbot/...) that contain full conversation history, tokens, and cost information — treat that as highly sensitive; only run the skill on a trusted machine and avoid exposing its output to untrusted parties; (2) it relies on common system tools (journalctl, systemctl, jq, grep, tail) though they are not declared as required—ensure those binaries exist and are the expected versions; (3) there is a minor metadata mismatch (ownerId differs in _meta.json vs registry metadata) — likely benign but worth noting. If you want extra assurance, inspect the session files yourself before running or run the scripts with limited privileges / under a non-privileged account.

Clawdbot Logs & Diagnostics

Analyze Clawdbot performance, errors, and session data.

Quick Commands

Response Times (last N messages)

scripts/response-times.sh [count]

Recent Errors

journalctl --user -u clawdbot-gateway.service --no-pager --since "1 hour ago" | grep -iE "(error|fail|invalid)" | tail -20

Session Stats

scripts/session-stats.sh

Gateway Status

systemctl --user status clawdbot-gateway.service --no-pager

Config Validation

cat ~/.clawdbot/clawdbot.json | jq . > /dev/null && echo "Config valid" || echo "Config invalid"

Log Sources

Source	Location	Contains
Journal	journalctl --user -u clawdbot-gateway.service	Session state, errors, tool exec
Daily log	/tmp/clawdbot/clawdbot-YYYY-MM-DD.log	Detailed JSON logs
Session file	~/.clawdbot/agents/main/sessions/*.jsonl	Full conversation, token usage, costs
Sessions meta	~/.clawdbot/agents/main/sessions/sessions.json	Current session state, model info

Common Diagnostics

Slow Responses

1. Check response times: scripts/response-times.sh 20
2. Check token count in sessions.json: jq '.["agent:main:main"].totalTokens' ~/.clawdbot/agents/main/sessions/sessions.json
3. If tokens > 30000, run /compact in Telegram or start new session

Config Errors

journalctl --user -u clawdbot-gateway.service --no-pager --since "10 minutes ago" | grep -i "invalid config"

API Costs (from session)

scripts/session-stats.sh

Useful Patterns

Filter journal by category

# Session state changes
journalctl --user -u clawdbot-gateway.service | grep "session state"

# Tool execution
journalctl --user -u clawdbot-gateway.service | grep "\[tools\]"

# Telegram activity
journalctl --user -u clawdbot-gateway.service | grep "\[telegram\]"

Parse session file for recent messages

tail -20 ~/.clawdbot/agents/main/sessions/*.jsonl | jq -r 'select(.message.role=="user") | .message.content[0].text' 2>/dev/null | tail -10