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 sigilagent/sigil --skill writing-agirgit clone --depth 1 https://github.com/sigilagent/sigilWrote 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/sigilagent/sigil/writing-agir)<a href="https://agentmods.dev/skills/sigilagent/sigil/writing-agir"><img src="https://agentmods.dev/badge/skills/sigilagent/sigil/writing-agir/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/sigilagent/sigil/writing-agir"><img src="https://agentmods.dev/badge/skills/sigilagent/sigil/writing-agir.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 | $0.00048 | $0.01488 |
| Opus 5 | $0.00024 | $0.00744 |
| Sonnet 5 | $0.00010 | $0.00298 |
| Haiku 4.5 | $0.00005 | $0.00149 |
Grade A, and why
writing-agir 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 9d 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 — 118 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Writing an AG-IR
You are the front-end of a compiler. Your input is a SKILL.md written for a
human, plus a frozen rule set already extracted from it. Your output is one
file, agent.ir: a YAML AG-IR that Sigil lowers mechanically into a runnable
object-spatial program.
The point of the AG-IR is that it removes the model's discretion. In a prompt-based harness a skill is advice the model may follow. Here every mandatory step becomes a node the run must visit, every prohibition becomes a constraint on a node, and every verification becomes a gate with a typed verdict. Your job is to build that structure faithfully — not to improve the skill, and not to summarize it.
The three things that get checked
Your draft is not accepted because it looks right. It is accepted because it clears three gates, and each one maps to a rule you must follow while drafting.
G4 — it must compile. sigil gate agent.ir <name> lowers your IR and
type-checks the result. Run it after every edit. It exits nonzero while the IR is
broken and prints the exact diagnostic.
G1 — every tool and knowledge item must be embodied. A tool with no body: and
no command: is a prose-pointer, and it is rejected. Knowledge must carry its
content verbatim, never "see the section above" or "follow the format in
FORMS.md". The test: if SKILL.md were deleted, would the IR still lower to a
working agent? If the answer depends on prose that only lives in the skill, embody
that prose in the IR.
G5 — every mandatory rule must be realized by a node. Each rule in the frozen
set has an id (r1, r2, …). Each node carries traces_to: [r3, r7] naming the
rules it realizes. A mandatory STEP or IO rule with no node tracing to it is a
structural gap and fails. Worse than missing is monolithic: several rules folded
into one model slot, so nothing guarantees each was done. Give a rule its own node
when it is its own obligation.
Prefer owner: code where the rule is mechanical — a code node is guaranteed by
construction. A model slot is only observable. Reserve model slots for genuine
judgment.
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.
- 9d ago First seen · 118 lines · 48 tokens per session scan A d2b8c467597f
writing-agir is a skill published in the GitHub repository sigilagent/sigil (11 stars, last pushed 9d ago), licensed MIT. It adds 48 tokens to every session and 1,488 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…
next-partial-prefetching-adoption
Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…