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 ulises-jeremias/dotfiles --skill hornerogit clone --depth 1 https://github.com/ulises-jeremias/dotfilesWrote 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/ulises-jeremias/dotfiles/hornero)<a href="https://agentmods.dev/skills/ulises-jeremias/dotfiles/hornero"><img src="https://agentmods.dev/badge/skills/ulises-jeremias/dotfiles/hornero/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/ulises-jeremias/dotfiles/hornero"><img src="https://agentmods.dev/badge/skills/ulises-jeremias/dotfiles/hornero.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00031 | $0.01045 |
| Opus 5.5 | $0.00012 | $0.00418 |
| Sonnet 5.5 | $0.00006 | $0.00209 |
| Haiku 4.5 | $0.00003 | $0.00104 |
Grade A, and why
hornero 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 2d 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 — 97 lines — stays where its author put it; the contents beside it link to each section on GitHub.
HorneroOS system skill
You are running on HorneroOS, an Arch-based Wayland desktop.
The canonical system interface is horneroctl (on PATH).
Prefer it over probing the filesystem or guessing paths: it is
faster, safer, and returns structured output with --json.
Golden rules
- Read before you act. Answer questions about this machine
from
horneroctloutput, not from generic Linux knowledge. - Read-only first. The commands below never change the system. Run them freely to build an answer.
- Mutations need the human. Anything that installs, removes,
deletes, writes config, or changes appearance must go through
the normal approval flow. Never add
--yesyourself, never auto-answer an approval, and never run a command the user has not seen. If an approval is pending, say so and stop. - Structured arguments. Call commands as literal argv (e.g.
horneroctl system info --json), never via string-built shell pipelines you invented, and never withsudounless the user explicitly asked for a privileged operation and approved it. - No secrets. Do not read or print API keys, tokens, or the
assistant key file (
~/.config/hornero/assistant.key).
Read-only tool catalogue
Discover the installed interface first with
horneroctl assistant tools --json; releases differ, so do not
assume a command exists. When you cannot execute commands at all,
ask the user (or a shell-capable step) for a fresh
horneroctl assistant knowledge --json snapshot — it is generated
from the same live allowlist with a date and provenance, so prefer
it over any copied table, including this one. This mirrors that
output (class: safe-read-only):
| Question | Command |
|---|---|
| System, kernel, hardware summary | horneroctl system info --json |
| Current appearance state | horneroctl appearance status --json |
| Available themes | horneroctl appearance theme list --json |
| One theme's details | horneroctl appearance theme show <name> --json |
| GTK themes installed | horneroctl appearance gtk list --json |
| Current GTK theme | horneroctl appearance gtk current --json |
| Pending package updates | horneroctl package updates --json |
| Installed plugins + state | horneroctl plugin list --json |
| Assistant status | horneroctl assistant status --json |
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.
- 2d ago Changed · +4 lines 98c037eaea14
- 3d ago First seen · 93 lines · 31 tokens per session scan A 41c310aaeda1
hornero is a skill published in the GitHub repository ulises-jeremias/dotfiles (135 stars, last pushed 2d ago), licensed MIT. It adds 31 tokens to every session and 1,045 once invoked, about $0.0001 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 2026-10-20.
Other skills, from other repositories
vale
Audit, write, or revise .vale.ini, or work through vale findings on a repo. Use when adding Vale, troubleshooting silent passes or noisy findings, evolving config, or applying the suggestion → warning → error cleanup cascade after a vale run (interactive or pre-commit failure). Applies the two-hook pre-commit pattern…
renovate
Audit, write, or revise renovate.json. Use when adding Renovate, troubleshooting unexpected (or missing) update PRs, hardening against supply-chain attacks, or evolving an existing config.
milestones
Create, scope, assign, score, and close GitHub milestones as monthly OKRs written to Google's OKR playbook — one committed or aspirational objective per milestone, key results as an "As measured by" checklist scored 0.0–1.0. Use when adding a milestone, writing or reviewing its objective and key results, deciding…
dfd
Audit, write, or revise a data flow diagram (DFD). Use when asked to draw or model how data moves through a system, when standing up a DFD with trust boundaries for threat modeling or security review, or when checking an existing DFD for drawing and balancing errors. Pins Mermaid flowchart as the output notation and…
pr-create
Open a GitHub PR with a tight body — why this change, what'll surprise the reviewer, how we know it's correct — or add a pullrequesttemplate to a repo. Use when asked to open, create, or draft a pull request, compose or regenerate a PR description, or scaffold a repo's PR template. Composes the…
pr-review
Review a pull request's diff against the cross-repo judgment bar — correctness, secrets, inline suppressions, matching the repo's conventions, docs, commits, and the test plan. Use when asked to review a PR or a branch diff for what the repo's linters, type checkers, and scanners can't assert. Assumes the sensors are…