analyze-log

name	analyze-log
description	Analyze a gptel HTTP log file and trace the multi-agent conversation flow
context	inline
user-invocable	true
model-invocable	false
argument-hint	<log-file-path>

Analyze the gptel HTTP log file at the path given in $ARGUMENTS.

If no path is given, look for the most recent *.log file in the current workspace root.

Run the analyzer script:

python3 ${CLAUDE_SKILL_DIR}/../../scripts/analyze-gptel-log.py $ARGUMENTS

After seeing the trace output, provide a brief interpretation covering:

1. Flow correctness: Did requests flow in the expected order? Did each agent get the right number of turns?
2. Result deliveries: Which agents received terminal RESULT envelopes, and when?
3. Message delivery: Were any SendMessage MAIL envelopes visible?
4. Anomalies: Any missing responses, unexpected agent counts, or mismatched token counts?
5. Timing: How long did the full flow take? Any suspiciously long gaps?

Keep the interpretation concise (under 200 words).