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 instructions/robcsaszar/tend/agents-mdgit clone --depth 1 https://github.com/robcsaszar/tendWhat 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 | $0.01098 | $0.01098 |
| Opus 5 | $0.00549 | $0.00549 |
| Sonnet 5 | $0.00220 | $0.00220 |
| Haiku 4.5 | $0.00110 | $0.00110 |
Grade A, and why
tend AGENTS.md 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 yesterday.
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 — 38 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Mission
This repo publishes the tend skill pack: local, on-demand maintenance skills for SvelteKit + TypeScript repos. There is no build, no tests, no runtime of its own. The deliverable is the contents of skills/*. Every skill shares one shape (config → triage → scan → fix → verify, one atomic change per run, never commit) and one shared config file (.claude/tend/config.yaml, schema owned by tend-onboard). Changes should be judged by: would a stranger who installs one of these skills into their own SvelteKit repo get a correct, safe, minimal fix out of it?
Layout convention
- One directory per skill under
skills/<name>/,name:in frontmatter matching the directory name exactly. - Every skill follows the same phase structure:
0. Load config→1. Triage(MANDATORY READ a core reference) →2. Scan(stop at first real hit; module-gated checks load only when their module is active) →3. Fix(one atomic, verified change) →4. Verify & present(re-run checks, show diff, never commit) →## NEVER(3-line**NEVER**/**Instead:**/**Why:**format). Keep new skills or edits to existing ones in this shape: it's load-bearing, not decorative;tend-onboard's config schema and every other skill's Phase 0 wording assume it. - Structured-findings skills (
tend-security,tend-refactor,tend-tests) shipscripts/validate-*.mjs+assets/*.json. Diff-is-the-finding skills (tend-perf,tend-a11y,tend-docs) don't. Don't add a validator to one of these without updating this file's reasoning for why not. - Every skill ships
evals/eval-N/{prompt,assertions}.md. Seeai-forge-evalin the source repo (orakl) for the assertion format if extending. - Every skill sets
disable-model-invocation: truein frontmatter: user-invoked only (/tend-security, etc.), never auto-triggered from conversation. This is the design decision behind the whole pack (single-skill, on-demand, human-paced runs replacing an unattended CI fleet that produced an 84-PR backlog). A skill that auto-triggers reintroduces the unattended-breadth failure mode the pack exists to avoid. New skills must set this flag too; don't add trigger-phrase language to a description once this flag is set, it's dead weight since auto-matching never runs. skills/tend-onboard/assets/tend-sweep.ymlis a copy-out template for consuming repos, installed (opt-in, cadence-adjusted) bytend-onboard's Phase 4..github/workflows/in this repo stays empty by design — never turn the template into a live workflow here.
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.
- yesterday First seen · 38 lines · 1,098 tokens per session scan A b0d7e7969d6e
tend AGENTS.md is an instructions file published in the GitHub repository robcsaszar/tend (0 stars, last pushed 2d ago), licensed MIT. It adds 1,098 tokens to every session, about $0.0055 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 2026-08-31.
Other instructions, from other repositories
agentic-hardening-skill AGENTS.md
Instructions for Calvin-LLC/agentic-hardening-skill, covering harden and rules.
craft-skills CLAUDE.md
Instructions for mikestangdevs/craft-skills, covering claude.md, what this repo is, commands, architecture and bucket organization & governance.
Codex-Skills AGENTS.md
Instructions for TheGoat395/Codex-Skills, covering repository operating notes, required checks, evidence rules and change hygiene.
dxkit CLAUDE.md
Instructions for vyuh-labs/dxkit, covering claude.md — dxkit development rules, architecture rules, 1. tool invocation goes through the registry, 2. never duplicate tool invocation logic and 3. language facts come from detect.ts.
humane-agentic-design CLAUDE.md
Instructions for glebis/humane-agentic-design, covering claude.md, what this repository is, rule ownership, what we deliberately do not cover and authoring conventions.
OpenGUI CLAUDE.md
Claude Code instructions for Core-Mate/OpenGUI, covering opengui release - repository guide for claude code / codex, 1. repository positioning, 2. default working rules, 3. high-level architecture and 4. directory map.