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 yqi96/warranted --skill literature-writinggit clone --depth 1 https://github.com/yqi96/warrantedWrote 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/yqi96/warranted/literature-writing)<a href="https://agentmods.dev/skills/yqi96/warranted/literature-writing"><img src="https://agentmods.dev/badge/skills/yqi96/warranted/literature-writing/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/yqi96/warranted/literature-writing"><img src="https://agentmods.dev/badge/skills/yqi96/warranted/literature-writing.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 | $0.00055 | $0.01669 |
| Opus 5.5 | $0.00022 | $0.00668 |
| Sonnet 5.5 | $0.00011 | $0.00334 |
| Haiku 4.5 | $0.00006 | $0.00167 |
Grade A, and why
literature-writing 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 1mo 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 — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Graph Mapping
Never cite a paper directly. A citable finding from a paper first becomes a proposition with that paper in its evidence.attachments; LaTeX then cites that proposition with \cite{prop_N}, where N is its id. The citation is a pointer into the argument graph, not a bibliography shortcut.
| What you are writing | Where it goes |
|---|---|
| citable finding from a paper | a proposition — the paper in evidence.attachments |
| paragraph point | a proposition — the literature propositions in its evidence.nodes |
| why those findings support the point | that proposition's warrant |
| methodological authority behind the inference | evidence on the warrant — promote_warrant first, then attach it there |
| conflicting paper finding | a proposition created with attacks={node, slot} |
There is no separate node type for any of these. A proposition is a ground when something else cites it and a claim when it cites something else, and the same node is usually both — the role is the viewpoint, not a property to declare.
Granularity:
- one paper yields multiple propositions when it offers multiple distinct citable findings
- one proposition carries multiple paper attachments when several papers support the same finding
- different findings, populations, methods, measurements, or scopes are separate propositions
- convergence across distinct propositions is the consuming proposition's
warrant, not a merge into a vague composite
Obligations
Object-layer work in this channel — search, reading, download, BibTeX, prose edits — is valid only when it creates, judges, cites, or reconciles a proposition the manuscript rests on. A searched paper must become evidence for a proposition or a promoted warrant, support a rebuttal proposition, or be discarded with a reason; it is never accumulated as an inert reading list.
| Graph state | Required action |
|---|---|
| No proposition for a load-bearing paragraph thesis | create_propositions, or update the existing one |
| A sentence cites a paper for a finding | Reuse a matching proposition, or create_propositions with the paper attached |
| A paper provides several relevant findings | One item per distinct finding in the same create_propositions call |
| Several papers support the same specific finding | Attach them to the same proposition only when they support the same finding |
| A cited proposition's paper has never been checked against its content | Read the paper, then update_proposition(id=<id>, evidence={add_attachments: [...]}) and judge it with set_qualifier(updates=[{id: <id>, qualifier: <band>}]) |
A paragraph thesis has evidence but no warrant |
update_proposition(id=<id>, warrant=...) — the structural check demands it from possibly upward |
| The warrant itself needs authority, or is what a source disputes | promote_warrant(id=<id>), then attach evidence to the promoted proposition |
| A paper contradicts a thesis or its warrant | create_propositions(items=[{content: <finding>, attacks: {node: <target id>, slot: <content or warrant>}}]) |
| The text changes the paragraph thesis | Reconcile the proposition's content |
| A citation points to the wrong proposition | Point to the correct one, or create one; edit an existing proposition only to fix an error, never to fit the sentence |
Existing LaTeX uses author-year or legacy statement_N keys |
Migrate each: identify the finding actually cited, create or reuse its proposition, replace the key with prop_N. Never map statement_7 to prop_7 by number — the old ids belong to an archived graph |
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.
- 1mo ago Changed · +4 lines · +2 tokens per session 426f66f98c80
- 1mo ago First seen · 88 lines · 53 tokens per session scan A 3ed5f9db52dc
literature-writing is a skill published in the GitHub repository yqi96/warranted (2 stars, last pushed 1mo ago), licensed MIT. It adds 55 tokens to every session and 1,669 once invoked, about $0.0002 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
proof-checker
A mathematical proof review and repair workflow for LaTeX documents. It checks whether a proof has valid reasoning, addresses identified gaps, reviews the fixes, and produces an audit report.
auto-paper-improvement-loop
An automated editing cycle for a compiled research paper: an external language model reviews it, fixes are applied, and the paper is compiled again.
paper-write
A workflow for writing a research paper section by section in LaTeX, a text format used to prepare structured technical documents. It starts from a paper outline and supports several academic venues and citation sources.
research-lit
Search and analyze research papers, find related work, summarize key ideas. Use when user says "find papers", "related work", "literature review", "what does this paper say", or needs to understand academic papers.
paper-illustration
A workflow for generating academic illustrations, such as architecture diagrams and method visuals, with image generation and repeated review. Claude plans and checks the figure during the process.
citation-crosscheck
Double-blind bibliographic cross-check: one agent extracts each cited entry verbatim from the local .bib, a second independently re-fetches the same paper's record from the web while blind to the .bib, and a third extractor emits a mechanical field-by-field delta to catch dropped/re-ordered authors, wrong years, wrong…