From claude-3t
Show current 3-tier agent status — hot memory state, cold-storage index, model configuration, approval queue, and session overview. Run anytime to get a full picture of where things stand.
How this skill is triggered — by the user, by Claude, or both
Slash command
/claude-3t:3t-statusThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Generate a full status report. Check each item and report honestly.
Generate a full status report. Check each item and report honestly.
Read each memory file and report:
Format as:
HOT MEMORY STATUS
─────────────────────────────────────────────
MEMORY.md Last active: [date]
In progress: [brief]
Next action: [brief]
CONTEXT.md [N] domain terms | [N] ADRs
EXECUTOR_MEMORY.md [N] entries | Latest: [task type]
IMPLEMENTOR_MEMORY.md [N] entries | Latest: [cause]
OVERRIDE_LOG.md [N] overrides this session
Pre-approved: [list or: none]
Read .claude/context/cold/INDEX.md and report:
COLD STORAGE
─────────────────────────────────────────────
Index entries: [N]
Topics: [comma-separated list of cold file topics, or: none]
Orphan check: [any cold/*.md file with no INDEX row? list, or: clean]
The orphan check matters — an unindexed cold file is invisible. If found, flag it: the INDEX must be updated or the file removed.
MODEL STATUS
─────────────────────────────────────────────
Executor: the model selected for this session (your session model)
Implementor: Haiku, latest (claude-3t:implementor subagent)
Reasoning: Opus via advisor (architect tier, automatic, no artifacts)
opusplan: [active if Shift+Tab plan mode configured | not configured]
Run /status for live quota consumption.
Check MEMORY.md Current State for any queued approvals:
APPROVAL QUEUE
─────────────────────────────────────────────
[list queued tasks or: none pending]
Provide a one-paragraph summary of what has been accomplished this session and what is recommended next.
Review this session for protocol violations and report honestly:
PROTOCOL COMPLIANCE
─────────────────────────────────────────────────────
PRE-AGENT CHECKLIST shown before agent calls? [yes | no | no agents invoked]
Cold index handed to agents that needed it? [yes | n/a]
Hot memory loaded at session start? [yes | no]
ADRs authored by implementor (not executor)? [yes | n/a]
Extended context overrides logged? [yes | N/A | violation]
─────────────────────────────────────────────────────
[list any violations found, or: none detected]
If violations are found, show them explicitly. Do not minimise.
End with: "Run /claude-3t:3t-checkpoint to save current state. Run /claude-3t:3t-leaving before stepping away."
npx claudepluginhub jimcline/claude-3t --plugin claude-3tCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.