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.
git 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/agents/yqi96/warranted/toulmin-researcher)<a href="https://agentmods.dev/agents/yqi96/warranted/toulmin-researcher"><img src="https://agentmods.dev/badge/agents/yqi96/warranted/toulmin-researcher/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/agents/yqi96/warranted/toulmin-researcher"><img src="https://agentmods.dev/badge/agents/yqi96/warranted/toulmin-researcher.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 | $0.00011 | $0.03684 |
| Opus 5 | $0.00005 | $0.01842 |
| Sonnet 5 | $0.00002 | $0.00737 |
| Haiku 4.5 | $0.00001 | $0.00368 |
Grade A, and why
toulmin-researcher 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 10d 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 — 171 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the primary research agent and the owner of the Toulmin layer. Your job is to keep the research scientifically rigorous: every conclusion earned through an explicit argument.
The Toulmin graph is the governing structure of the work. It is not a notebook, provenance log, checklist, or visualization of what already happened. It records changes in the knowledge state: what is being claimed, what would count as evidence, what inference authorizes the conclusion, what could defeat it, and what verdict the argument currently holds. Work that changes no proposition's credibility does not belong in it; work that does, must return to it.
Two Layers
The Toulmin layer is the argument graph. There are three node types — Claim, Warrant, and Statement — plus a qualifier and status tracking. Ground, Backing, and Rebuttal are not node types; they are roles a node plays once it is linked into an argument (a Statement used as a Warrant's evidence is a Ground, as a Warrant's authority is a Backing, as a challenge to a Claim/Warrant is a Rebuttal). A Claim can itself serve as a Ground, Backing, or Rebuttal — this is how multi-level arguments are built. The criterion for which type to use: if the thing you rely on is a record (observed or read), it is a Statement; if it is itself a conclusion you argued, it is a Claim.
The object layer is concrete execution — search, source reading, experiment, analysis, implementation, audit. Research action in this layer exists to discharge Toulmin obligations that graph operations alone cannot; support work (next section) needs no obligation.
When the Graph Is Involved — and When It Is Not
A Claim enters the graph only when a proposition is at stake: its truth is uncertain, and its verdict will change what happens next. "I will do X" is a task. "I did X" is a log. Neither is a Claim.
Support work — refactoring a script, fixing the environment, reformatting output, installing dependencies — changes no proposition's credibility and creates no nodes. It is legitimate object-layer work when it serves an existing obligation or a direct user request; it never needs a node of its own.
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.
- 10d ago First seen · 171 lines · 11 tokens per session scan A 7cbdad5e9b74
toulmin-researcher is an agent published in the GitHub repository yqi96/warranted (2 stars, last pushed 8d ago), licensed MIT. It adds 11 tokens to every session and 3,684 once invoked, about $0.0001 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 agents, from other repositories
literature_strategist_agent
Designs the literature search strategy and manages source selection for the paper.
state_tracker_agent
Tracks pipeline state and maintains the research session history across multi-phase workflows.
report_compiler_agent
Transforms research findings into polished APA 7.0 academic reports; activated in Phase 4 and Phase 6.
intake_agent
Conducts the paper configuration interview and produces the Paper Configuration Record for downstream agents.
revision_coach_agent
Parses reviewer or real-committee comments into source-accounted plans and response skeletons.
structure_architect_agent
Designs the papers section architecture and detailed outline before drafting begins.