Stats
Actions
Tags
From agentguard
Diagnose a Claude session with AgentGuard. If no session id is given, list recent sessions for the current project and ask the user to pick one.
How this skill is triggered — by the user, by Claude, or both
Slash command
/agentguard:diagnose-sessionThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
`$ARGUMENTS` is an optional Claude session id.
$ARGUMENTS is an optional Claude session id.
Run and paste the output verbatim, then ask the user to pick one:
${CLAUDE_PLUGIN_ROOT}/bin/agentguard list-claude-sessions --limit 10 --project .
Wait for the user's choice. Do NOT auto-pick. If no sessions are listed, offer to run the list-sessions skill to look across all projects.
Run and paste the output verbatim:
${CLAUDE_PLUGIN_ROOT}/bin/agentguard diagnose-claude-session "$ARGUMENTS" --output .agentguard/traces/"$ARGUMENTS".json
The [artifacts] block in the output lists a trace JSON and a companion HTML report that is generated automatically. Do not add any summary unless the user asks.
npx claudepluginhub betahi/agentguard --plugin agentguardCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.