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 lukastk/boxyard --skill boxyard-cligit clone --depth 1 https://github.com/lukastk/boxyardWrote 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/lukastk/boxyard/boxyard-cli)<a href="https://agentmods.dev/skills/lukastk/boxyard/boxyard-cli"><img src="https://agentmods.dev/badge/skills/lukastk/boxyard/boxyard-cli/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/lukastk/boxyard/boxyard-cli"><img src="https://agentmods.dev/badge/skills/lukastk/boxyard/boxyard-cli.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- Socket pass
- Snyk 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 | $0.00064 | $0.01652 |
| Opus 5.5 | $0.00026 | $0.00661 |
| Sonnet 5.5 | $0.00013 | $0.00330 |
| Haiku 4.5 | $0.00006 | $0.00165 |
Grade A, and why
boxyard-cli 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 24d 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 — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Boxyard CLI
Use this skill to use, not develop, Boxyard: managed folders with DATA, metadata, sync config, groups, and local/rclone storage. All bundled paths below resolve from this skill directory; no source checkout is required.
Authority and safety
- Ask before modifying local or remote state. This includes
init,new,sync,multi-sync,sync-missing-meta,include,exclude,relocate,delete,rename,sync-name, group/parent edits,create-user-symlinks,copy,force-push,claim,release,discard-local, and non-dry-runconvert. Examples are not authorization. Read-only queries,--help,doctor, andconvert --dry-runmay run without confirmation. - Never create ad-hoc folders/clones/worktrees in
~/dev. Find the existing box or useboxyard newwith approval; put non-box scratch outside managed roots. Keep generated data, even large outputs, inside its related box. Use sync excludes, not sibling folders, to avoid uploading it. - Resolve identity and placement, don't guess. Both timestamp format and subid
length are configurable;
index_nameis{box_id}__{name}. Usewhich/path: DATA may be in any checkout root, not justuser_boxes_path(default). An unavailable root never falls back to default. Placement is machine-local, never synced inboxmeta.toml; remote storage and local placement are independent. - Resolve ownership, never bypass it. DATA/CONF pushes compare
write_ownerwith configuredmachine_name, never a hostname. No owner means unrestricted.newclaims by default (--no-claimfor work elsewhere);include --read-onlysuppresses the claim nudge. Refusal also protectsforce-push, remote/both rename, and deletion. With approval,claim --stealkeeps this machine's work;discard-localreplaces it from remote and keeps overwritten work in backups. - Destructive traps:
new --from PATHMOVES unless--copy;excludesyncs first then removes local DATA (--skip-syncskips that protection).deletetombstones the id fleet-wide: never delete a stale duplicate name to clean registrations.sync --sync-setting replace|force,copy --overwrite, andforce-push --forcecan overwrite data. Read the relevant reference first. - Backups aren't garbage: orphan findings may be intentional
discard-localkeepsakes. Never auto-delete them or infer disposability from unclaimed ULIDs. - Excludes replace, not extend. A box's
conf/.rclone_excludereplaces the default list entirely: copy the default contents before adding rules, or risk syncing.venv/,node_modules/, etc..rclone_includeand.rclone_filtersare currently refused; use.rclone_excludealone. CONF syncs before DATA. Default-exclude edits invalidate DATA baselines fleet-wide; batch needed edits. - Conversion is explicit: policy edits never reformat existing boxes; only
convertdoes, verifying a byte-identical restore before removing the old copy. Read the storage reference and~/.config/myagent/reference/box-storage.mdbefore recommending/changing plain vs restic. Rig default is plain; package default for rclone is restic (local is plain). Every participating machine needs Boxyard >=0.7.0, restic and the password before using restic boxes.
What ships with it
5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 24d ago Changed · -429 lines 39ba08752c34
- 1mo ago Changed · +58 lines 6abef1ff784e
- 2mo ago First seen · 488 lines · 64 tokens per session scan A fb057c01b2c6
boxyard-cli is a skill published in the GitHub repository lukastk/boxyard (0 stars, last pushed 25d ago), licensed MIT. It adds 64 tokens to every session and 1,652 once invoked, about $0.0003 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 2026-08-31.
Other skills, from other repositories
orbit-notion
Open Orbit briefing skill — selected by the Orbit pipeline when Notion is the user's only connected connector, or when the user explicitly scopes their daily digest to Notion. Pulls the past 24 hours of document edits, comments, mentions, and database row changes from the user's authenticated Notion connection and…
agentmail
Use your assigned AgentMail inbox to read email tasks, explicitly send or reply, and check delivery. Provided automatically by your inbox assignment.
rowboat-spaces
How to work as a member of Rowboat Spaces, the team chat where people and agents work together. Load it once per session, before acting on any request that reaches you from Rowboat; it covers your setup, mentions and hand-offs, the tools, and files.
Cortex
LifeOS memory, prior-session recall, and learning capture. USE WHEN capture this learning, document this learning, what did we learn recently, recent learnings, cortex, knowledge, knowledge base, search knowledge, what do we know about, archive, harvest, knowledge status, develop note, add to knowledge, add a people…
feishu
Work with Feishu or Lark bots, docs, sheets, bitables, approval flows, and OpenAPI/MCP setup without hardcoding credentials.
pinchtab-mcp
Use this skill when a task requires browser automation through PinchTab's MCP server connected to a remote browser instance. Covers navigation, element interaction, data extraction, form filling, multi-step flows, and session management via MCP tools.