How do I install the Tma1 Setup skill?

Run npx skillmds@latest add tma1-ai/tma1-setup in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

What does the Tma1 Setup skill do?

Install and configure TMA1 local observability (Claude Code, Codex, GitHub Copilot CLI, OpenClaw, any OTel SDK). Use when the user says: install tma1, setup observability, monitor my agent, track token usage, set up telemetry. It is listed under AI & ML on SkillMD.

Is Tma1 Setup safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Tma1 Setup?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Tma1 Setup free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Tma1 Setup?

tma1-ai (@tma1-ai) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Tma1 Setup by tma1-ai — Agent Skill

Name: Tma1 Setup
Author: tma1-ai

TMA1 Setup

You are helping the user install and configure TMA1, a local-first LLM agent observability tool.

Step 1: Check if TMA1 is already running

curl -sf http://localhost:14318/health

If this returns {"status":"ok"}, TMA1 is already running. Skip to Step 4.

Step 2: Install TMA1

Download and install the tma1-server binary:

# macOS / Linux
curl -fsSL https://tma1.ai/install.sh | bash

# Windows (PowerShell)
irm https://tma1.ai/install.ps1 | iex

This installs the binary to ~/.tma1/bin/tma1-server (or %USERPROFILE%\.tma1\bin\tma1-server.exe on Windows).

Step 3: Start TMA1

On macOS/Linux the installer registers a service that auto-starts. On Windows the installer registers a Scheduled Task that auto-starts.

If TMA1 is not running, start it manually:

# macOS / Linux
~/.tma1/bin/tma1-server &

# Windows (PowerShell)
Start-Process "$env:USERPROFILE\.tma1\bin\tma1-server.exe"

Wait for GreptimeDB to become healthy:

# macOS / Linux
for i in $(seq 1 30); do
  if curl -sf http://localhost:14318/health > /dev/null 2>&1; then
    echo "TMA1 is ready."
    break
  fi
  sleep 1
done

# Windows (PowerShell)
for ($i = 0; $i -lt 30; $i++) {
  try { if ((Invoke-WebRequest -Uri http://localhost:14318/health -UseBasicParsing).StatusCode -eq 200) { Write-Host "TMA1 is ready."; break } } catch {}
  Start-Sleep -Seconds 1
}

If it does not become healthy after 30 seconds, check the logs and report the error.

Step 4: Verify OTel endpoint

Confirm GreptimeDB is accepting OTLP data:

curl -sf http://localhost:14318/status

This should return {"status":"ok","greptimedb":"running",...}.

Step 5: Configure the agent

Tell the user to set the OTel exporter endpoint. The exact method depends on their agent:

Claude Code — merge into ~/.claude/settings.json (Windows: %USERPROFILE%\.claude\settings.json):

CRITICAL: You MUST read the existing settings.json first and MERGE — NEVER overwrite.

For "env": add/update only the keys shown below. Keep all existing env vars intact.

For "hooks": for each event type, append the TMA1 hook entry to the existing array. Do NOT replace the array or remove other hooks.

For all other top-level keys (permissions, mcpServers, enabledPlugins, etc.): do NOT touch them.

Example merge for a hook event that already has entries:
"PreToolUse": [
  { "hooks": [{ "type": "command", "command": "existing-hook.sh" }] },
  { "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }
]

The following keys need to be present (add if missing, do not remove others):

{
  "env": {
    "CLAUDE_CODE_ENABLE_TELEMETRY": "1",
    "CLAUDE_CODE_ENHANCED_TELEMETRY_BETA": "1",
    "OTEL_EXPORTER_OTLP_ENDPOINT": "http://localhost:14318/v1/otlp",
    "OTEL_EXPORTER_OTLP_PROTOCOL": "http/protobuf",
    "OTEL_METRICS_EXPORTER": "otlp",
    "OTEL_LOGS_EXPORTER": "otlp",
    "OTEL_TRACES_EXPORTER": "otlp"
  },
  "hooks": {
    "SessionStart": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "SessionEnd": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "PreToolUse": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "PostToolUse": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "PostToolUseFailure": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "UserPromptSubmit": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "SubagentStart": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "SubagentStop": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "Notification": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "Stop": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "PreCompact": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "PostCompact": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "PermissionRequest": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "PermissionDenied": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "TaskCreated": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "TaskCompleted": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "FileChanged": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "CwdChanged": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "InstructionsLoaded": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "Elicitation": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "ElicitationResult": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "WorktreeCreate": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "WorktreeRemove": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "StopFailure": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "Setup": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "TeammateIdle": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }],
    "ConfigChange": [{ "hooks": [{ "type": "command", "command": "~/.tma1/hooks/tma1-hook.sh", "timeout": 3 }] }]
  }
}

Claude Code exports metrics, logs, and traces (enhanced telemetry). CLAUDE_CODE_ENHANCED_TELEMETRY_BETA=1 enables trace spans with TTFT, tool timing, and permission waits for the Traces tab and waterfall. The hooks use command hooks (via ~/.tma1/hooks/tma1-hook.sh, auto-installed by tma1-server on startup) for all 27 event types. On Windows, use %USERPROFILE%\.tma1\hooks\tma1-hook.ps1 instead.

Codex — add to ~/.codex/config.toml (Windows: %USERPROFILE%\.codex\config.toml):

[otel]
log_user_prompt = true

[otel.exporter.otlp-http]
endpoint = "http://localhost:14318/v1/logs"
protocol = "binary"

[otel.trace_exporter.otlp-http]
endpoint = "http://localhost:14318/v1/traces"
protocol = "binary"

[otel.metrics_exporter.otlp-http]
endpoint = "http://localhost:14318/v1/metrics"
protocol = "binary"

Codex uses separate exporters for logs, traces, and metrics. Restart Codex after config changes.

OpenClaw:

openclaw config set diagnostics.enabled true
openclaw config set diagnostics.otel.enabled true
openclaw config set diagnostics.otel.endpoint http://localhost:14318/v1/otlp
openclaw config set diagnostics.otel.traces true
openclaw config set diagnostics.otel.metrics true
openclaw gateway restart

Session transcripts at ~/.openclaw/agents/*/sessions/*.jsonl are auto-discovered — no extra setup needed for Sessions and Prompts views.

GitHub Copilot CLI — zero config. TMA1 auto-discovers session logs at ~/.copilot/session-state/<sessionId>/events.jsonl. Nothing to configure; just run tma1-server and use Copilot CLI normally. Sessions, tool calls (with failure detection), subagent lifecycle, and skill invocations all show up in the Copilot CLI dashboard view and the unified Sessions view.

Other OTel-compatible agents (standard GenAI SDK) — typically export traces:

# macOS / Linux
export OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:14318/v1/otlp
export OTEL_EXPORTER_OTLP_PROTOCOL=http/protobuf

# Windows (PowerShell)
$env:OTEL_EXPORTER_OTLP_ENDPOINT = "http://localhost:14318/v1/otlp"
$env:OTEL_EXPORTER_OTLP_PROTOCOL = "http/protobuf"

Step 6: Verify data flow

After the user runs at least one agent interaction with the endpoint configured:

curl -s -X POST http://localhost:14318/api/query \
  -H 'Content-Type: application/json' \
  -d '{"sql": "SHOW TABLES"}' 2>/dev/null | python3 -m json.tool

If you see opentelemetry_logs, opentelemetry_traces, openclaw_*, claude_code_*, codex_*, tma1_hook_events, or tma1_messages tables, data is flowing. For Copilot CLI specifically, check that tma1_hook_events has rows with agent_source = 'copilot_cli'.

Handoff

Tell the user:

Dashboard: http://localhost:14318

Query API — all SQL queries go through POST with JSON body:
  curl -s -X POST http://localhost:14318/api/query \
    -H 'Content-Type: application/json' \
    -d '{"sql": "SHOW TABLES"}'

OTel endpoint: http://localhost:14318/v1/otlp
Use /tma1 to query observability data inline without opening the dashboard.
For more queries: https://tma1.ai/SKILL.md

Persistent processes — wrap with `tma1 build --watch`

For commands that run indefinitely until the user stops them (npm run dev, cargo watch, pytest --watch, hugo serve, docker compose up), wrap with tma1 build --watch:

tma1 build --watch --tag dev -- npm run dev
tma1 build --watch --tag watch -- cargo watch -x test

For one-shot commands (make test, go test ./..., pytest, cargo build), don't bother — just use the Bash tool directly. You already get the full output inline in the tool result, and any failure is visible to you immediately.

Why `tma1 build --watch` is worth it for persistent processes

The Bash tool's run_in_background mode can keep a dev server alive, but its output is fragmented across BashOutput polls and disappears when the session ends. tma1 build --watch solves that:

Persistence beyond your session — output is written to tma1_build_events; the next agent session (or another agent / the human at the dashboard) reads it via the get_build_status MCP tool without needing to be attached to your original process.
Time-debounced flush + signal forwarding — flushes buffered output every 2s (override with --debounce) instead of waiting for a line threshold, and forwards SIGINT/SIGTERM so Ctrl-C cleanly stops the dev server.
Anomaly detection — tma1_build_events feeds the repeated_failed_build and build_broken_after_my_edit rules, which surface in the next <tma1-context> injection if the dev server starts emitting failures.

Run tma1 build --help for the full flag list.

Tma1 Setup

File contents

TMA1 Setup

Step 1: Check if TMA1 is already running

Step 2: Install TMA1

Step 3: Start TMA1

Step 4: Verify OTel endpoint

Step 5: Configure the agent

Step 6: Verify data flow

Handoff

Persistent processes — wrap with `tma1 build --watch`

Why `tma1 build --watch` is worth it for persistent processes

Frequently asked questions

Tma1 Setup

File contents

TMA1 Setup

Step 1: Check if TMA1 is already running

Step 2: Install TMA1

Step 3: Start TMA1

Step 4: Verify OTel endpoint

Step 5: Configure the agent

Step 6: Verify data flow

Handoff

Persistent processes — wrap with tma1 build --watch

Why tma1 build --watch is worth it for persistent processes

Related skills

Frequently asked questions

How do I install the Tma1 Setup skill?

What does the Tma1 Setup skill do?

Is Tma1 Setup safe to use?

Which AI agents work with Tma1 Setup?

Is Tma1 Setup free to use?

Who published Tma1 Setup?

Persistent processes — wrap with `tma1 build --watch`

Why `tma1 build --watch` is worth it for persistent processes