Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/raylanlin/smarttune-cli/skill-mcpnpx skills add raylanlin/smarttune-cli --skill skill-mcpgit clone --depth 1 https://github.com/raylanlin/smarttune-cliWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00086 | $0.03050 |
| Opus 5 | $0.00043 | $0.01525 |
| Sonnet 5 | $0.00017 | $0.00610 |
| Haiku 4.5 | $0.00009 | $0.00305 |
Grade A, and why
smarttune-mcp scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 172 lines — stays where its author put it; the contents beside it link to each section on GitHub.
SmartTune MCP
Use this skill when SmartTune is connected through MCP and the agent does not have shell, exec, or write permission.
This is the MCP-only variant of the SmartTune skill. If the agent has normal CLI access, use the smarttune skill instead.
Hard Rules
- Do not call
exec, shell commands, Python subprocesses, or thestuneCLI. - Do not ask for write, edit, or process execution permissions to analyze a log.
- Do not create report files or choose arbitrary output paths.
- Do not modify logs, parameters, firmware, local files, or aircraft configuration.
- Use only the SmartTune MCP tools listed below.
- If these tools are unavailable, say SmartTune MCP is not connected and ask the operator to enable the MCP server.
- Analysis results arrive pre-validated (v3.2.1). Every recommendation returned by
smarttune_analyze_logand the per-module analysis tools carriesvalidated/validation_status. Only recommendations withvalidated: truemay be presented as-is; afalseentry must be dropped or corrected using theoptionsfromsmarttune_get_param. - MANDATORY: Validate every value YOU adjusted. Before suggesting any parameter change, call
smarttune_validate_paramwith the exact parameter name, proposed value, and platform. Never recommend a parameter that fails validation — search for alternatives withsmarttune_search_params. Averdictofunverifiableis not approval. For a whole set of adjusted values, usesmarttune_validate_params(one call instead of N). - Every tool returns one shape:
{ok: true, ...}on success,{ok: false, error_code, message, hint, retryable}on failure. A rejected parameter value is a successful call withvalid: false. - Firmware-version tables (v3.3). All six parameter tools accept
fw_version(e.g."copter-4.5"); omit it for the platform default (ArduPilot default = Copter-4.1). An unknown version returnserror_code: "E4011"with the available list. Ranges can differ between versions — validate against the version the user actually flies. - Never list a whole parameter table. Browse with
smarttune_list_param_groups→smarttune_list_params(group=...)→smarttune_get_param(name)for full detail. ArduPilot alone is ~2,800 parameters.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 172 lines · 86 tokens per session scan A 759f18547eba
smarttune-mcp is a skill published in the GitHub repository raylanlin/smarttune-cli (28 stars, last pushed 20d ago), licensed MIT. It adds 86 tokens to every session and 3,050 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
exporting-visualizations
Use when the user wants to SEE robot data — visualize, plot, replay, or export a log or query result from bagel to a viewer or training format (Rerun, Lichtblick, PlotJuggler, LeRobot).
operating-live-sinks
Use when subscribing to live robot data (MQTT, rosbridge) through bagel, managing standing pipelines that survive restarts, or when questions about live-buffer sizes and disk usage come up.
triaging-robot-logs
Use when investigating a robot log or flight log with bagel — finding events, anomalies, or "what happened around X" in rosbags, MCAP, CAN, MDF4, ULog, or Betaflight files. Routes to the server's triage capability.
authoring-pipelines
Use when the user wants a bagel data pipeline — reducing a log to windows around events ("keep 30s around every hard brake"), recurring snippets/GIFs/exports, or any tasks-and-gates automation over robot data.
drone_commands
Execute drone commands -- the AIPass CLI interface for all module operations.
cloud-api
处理大疆上云 API（DJI Cloud API）的接入、解析、调试与问题定位。当用户提到"大疆上云""Cloud API""云平台对接""大疆机场上云""Pilot 上云""MQTT 设备接入""物模型""HMS 告警""机场直播""OSD 数据""DRC 指飞""航线任务下发""设备拓扑""固件升级"等涉及大疆行业设备（机场/DJI Pilot 2/遥控器/无人机）与第三方云平台通信的场景时使用。.