Burrow is a native Mac app for cleaning, uninstalling, analyzing, optimizing, and monitoring a computer, with an engine and an MCP server that lets AI agents interact with it. It is for Mac users and coding agents that need to inspect and maintain system storage and status. The catalogue skill helps AI agents use Burrow's agent-facing features.
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 skills add caezium/burrow --skill burrow-system-toolsgit clone --depth 1 https://github.com/caezium/burrowWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/caezium/burrow/burrow-system-tools)<a href="https://agentmods.dev/skills/caezium/burrow/burrow-system-tools"><img src="https://agentmods.dev/badge/skills/caezium/burrow/burrow-system-tools/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/caezium/burrow/burrow-system-tools"><img src="https://agentmods.dev/badge/skills/caezium/burrow/burrow-system-tools.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What 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 5.1 | $0.00194 | $0.02069 |
| Opus 5.5 | $0.00078 | $0.00828 |
| Sonnet 5.5 | $0.00039 | $0.00414 |
| Haiku 4.5 | $0.00019 | $0.00207 |
Grade A, and why
burrow-system-tools 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 1mo 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 — 130 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Burrow system tools
Burrow runs a local MCP server over the user's Mac: live + historical system state (read-only) and gated maintenance. The governing habit is diagnose first — when a question is about this machine, or you spot a system symptom mid-task, reach for the read-only tools, name the cause, then propose a fix. Read-only tools never change anything, so there's no reason to hesitate.
Diagnose first (read-only — always safe)
- burrow_doctor — one-call health sweep: engine present, Full Disk Access, memory pressure, disk headroom, SMART disk health, Time Machine backup age, recent decode errors. Start here for any vague "something's wrong / is my Mac healthy?" — it tells you which area to drill into. It does not report SIP / Gatekeeper / FileVault / firewall over MCP (only the GUI fills those in), so for "is my Mac secure?" read them from the shell rather than claiming the tool checked them.
- burrow_snapshot — current vitals (CPU, memory, disk, network, temperature, top processes, a 0–100 health score). For "what's happening right now".
- burrow_top_processes — top CPU right now. For "what's using my CPU / why is it hot or loud?"
- burrow_process_usage — ranks over a window by cpu_time / peak_cpu / avg_cpu / peak_mem. Prefer this for "all day / since this morning / what's draining my battery?"
- burrow_history / burrow_diff — a trend over time, or what changed since a point ("it got slow in the last hour").
- burrow_disk_forecast — "when will my disk fill up?" (pointless once the
disk is already full — go straight to analyze). burrow_analyze
<path> — "what's eating space in <folder>?" Supports
depth(descend into the largest subdirectories in one call),limit, andmin_size— e.g.depth: 2, min_size: 104857600maps hotspots without a call per directory. Scanning a home folder or ~/Library can take minutes: pass the most specific path you can. - burrow_ports — "what's listening / what's on port 3000?" (pid + owner).
- burrow_cleanup_history / burrow_deleted_files — what Burrow has cleaned, and exactly which files it removed.
- burrow_list_apps — installed apps + the exact names uninstall accepts (call this before any uninstall). burrow_info — whether Burrow is even recording data (use when results look empty or stale).
- Reclaim candidates (read-only, report-only — they find things worth
deleting but never delete): burrow_dupes
paths(duplicate files), burrow_photospath(visually near-duplicate images), burrow_orphanspath(files belonging to no installed app), burrow_sentinel (apps sitting in the Trash whose leftovers you could sweep), burrow_slim_checkbinary(how much thinning a fat binary would reclaim), burrow_net (which app is moving bytes right now), burrow_rules_dryrundir(what a community rules directory would target). - burrow_anomalies — processes whose last-24h CPU has regressed against their own 14-day baseline. Reach for it when the user says something feels off but nothing looks obviously high: this is per-process, so a program that always sits at 40% isn't flagged and one that went 2% → 15% is.
- burrow_agent_audit — what agents (including you, earlier) have already run through this server: the tool, the exact arguments, dry-run or real, and the outcome. Check it before repeating a cleanup, and whenever you're not sure a call went through.
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.
- 1mo ago First seen · 130 lines · 194 tokens per session scan A bd721d7d1f5e
burrow-system-tools is a skill published in the GitHub repository caezium/burrow (1,594 stars, last pushed today), licensed MIT. It adds 194 tokens to every session and 2,069 once invoked, about $0.0008 per session on Opus 5.5. 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
cli-anything-nslogger
CLI harness for NSLogger — parse, filter, export, and monitor NSLogger log files (.rawnsloggerdata / .nsloggerdata).
sentry-app-setup
Create and configure a Sentry internal integration so the assistant can manage issues, alerts, and releases under its own identity.
dustoff
Assess a project that's been dormant for months and produce a prioritized re-entry plan — lifecycle stage, what's stale or broken, and how far the project's tooling has drifted from current conventions — delegating every fix to the skill that owns it, and optionally capturing the plan as a tracking issue that survives…
qa-triage
Triage a QA-labeled report — investigate it against the code, classify it, and draft the technical issue(s) it warrants, stopping for approval before creating anything.
wispterm-diagnostics
Use when a user wants to report, troubleshoot, or collect context for a WispTerm issue, including crashes, rendering/DPI glitches, high CPU, keyboard/input bugs, selection/copy/scrolling, SSH/SCP failures, SSH image preview failures, HTML preview/browser panel failures, SSH disconnects such as…
process-hygiene
Report, and with --reap stop, orphaned processes that interrupted agent test runs left behind in per-session scratchpads (PPID 1, this user, older than 2 h by default, cwd inside /claude- / / /scratchpad, and that session has ended). --spotlight keeps Spotlight from indexing scratchpad churn. For Sutando developers…