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 dryvist/claude-code-plugins --skill claude-skill-authoringgit clone --depth 1 https://github.com/dryvist/claude-code-pluginsWrote 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/dryvist/claude-code-plugins/claude-skill-authoring)<a href="https://agentmods.dev/skills/dryvist/claude-code-plugins/claude-skill-authoring"><img src="https://agentmods.dev/badge/skills/dryvist/claude-code-plugins/claude-skill-authoring.svg" alt="Measured on agentmods" 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 | $0.00054 | $0.01246 |
| Opus 5 | $0.00027 | $0.00623 |
| Sonnet 5 | $0.00011 | $0.00249 |
| Haiku 4.5 | $0.00005 | $0.00125 |
Grade A, and why
claude-skill-authoring 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 — 125 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AgentsMD Authoring Standards
Commands are skills. A slash command and a skill are the same mechanism — a
SKILL.md (or a legacy flat commands/*.md) that you or the model can invoke.
Author everything new as a skill; flat commands/*.md still work but are legacy.
File Structure
agentsmd/ # Single source of truth
├── AGENTS.md # Main entry point
├── rules/ # Auto-load every session (via .claude/rules symlink)
├── skills/ # On-demand (via .claude/skills symlink)
├── agents/ # Task subagents (via .claude/agents symlink)
└── workflows/ # Development workflow docs
.copilot/, .claude/, .gemini/ # Vendor dirs — symlinks only, no duplicates
Token Budget
Everything is measured in tokens — the actual context cost — not lines or
bytes. Count with the same tool the guard uses: atc -m sonnet. The budget
depends on when the file loads into context:
| Tier | Files | Loads | Target | Max |
|---|---|---|---|---|
| Always-on | CLAUDE.md/AGENTS.md, .claude/rules/*, plugin rules/* |
every session | 500 | 1000 |
| On-demand body | skills/*/SKILL.md, agents/*.md, commands/*.md |
when invoked, then persists | 1500 | 3500 |
| Reference | skills/*/references/*.md |
only when the body links to it | single-topic | — |
A skill description is always loaded (it drives skill selection): keep the
combined description + when_to_use within the 1,536-char listing cap and
lead with the use case.
Canonical catalogs and state machines (e.g. gh-cli-patterns, finalize-pr)
may exceed the on-demand max — record an explicit waiver rather than splitting a
procedure that must stay whole.
The two rules that override token count
- Self-contained execution. A skill must run correctly and safely when it loads alone. Before trimming or moving anything out, ask: "if only this skill loaded, would it still work?" If not, the content stays inline. Token count is a flag, never a reason to break this.
- Safety reminders stay inline — even when duplicated. A "see X" reference is absent when only the referencing skill loads, so keep invocation-local correctness/safety content inline in every skill that acts on it (destructive git ops, merge-readiness gates, "don't act on stale state", least-privilege rules in editing agents). Deduplicate reference material (command catalogs, queries, examples, tables) — never decision logic a caller needs before it knows to load the source.
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 · 125 lines · 54 tokens per session scan A b655c9ee2f3f
claude-skill-authoring is a skill published in the GitHub repository dryvist/claude-code-plugins (3 stars, last pushed yesterday), licensed Apache-2.0. It adds 54 tokens to every session and 1,246 once invoked, about $0.0003 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 2026-09-04.
Other skills, from other repositories
android-development
Android development with Kotlin, Jetpack Compose, and modern Android architecture. Use when building Android apps, implementing Material Design, or following Android best practices.
ai-ml-development
AI and machine learning development with PyTorch, TensorFlow, and LLM integration. Use when building ML models, training pipelines, fine-tuning LLMs, or implementing AI features.
authentication-patterns
OAuth 2.0, JWT, SSO, MFA, NextAuth/Clerk/Supabase Auth implementation patterns.
email-systems
Transactional email (Resend, SendGrid, SES), templates (React Email, MJML), deliverability (SPF/DKIM/DMARC), and inboxing best practices. Use when building email infrastructure, designing templates, or troubleshooting deliverability.
customer-persona-builder
Data-driven customer persona development combining market research, user behavior analysis, and segmentation frameworks. Use when creating buyer personas, ideal customer profiles (ICPs), or user archetypes.
devops-cloud
DevOps, cloud infrastructure, and platform engineering. Use when working with AWS, GCP, Azure, Kubernetes, Terraform, CI/CD pipelines, or infrastructure as code.