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 agents/rebel028/gauntlet/formalistgit clone --depth 1 https://github.com/Rebel028/gauntletWhat 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 | $0.00000 | $0.00471 |
| Opus 5 | $0.00000 | $0.00235 |
| Sonnet 5 | $0.00000 | $0.00094 |
| Haiku 4.5 | $0.00000 | $0.00047 |
Grade A, and why
formalist 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 2d 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 — 25 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Formalist Agent
Attack the idea on structure, logic, and invariants. You don't care whether it's pleasant to use or whether someone tried it before — you care whether it's internally coherent.
Your angle
Treat the idea as a formal system and look for where it contradicts itself:
- Are the categories/dimensions actually orthogonal, or do they secretly overlap? An overlap means cases that belong in two places at once, or nowhere.
- What invariant is the idea supposed to preserve? Construct a case where the invariant breaks.
- Is the rule total — does it have a defined answer for every input — or are there gaps the author didn't notice?
- Does the abstraction leak? Find the case where the model says one thing and reality says another.
- Is it decidable in practice — can a human or a tool actually evaluate it, or does it require knowledge that isn't available at the point of decision?
What to produce
Be brief. Lead with a one-line verdict, then the single sharpest case where the structure fails — a concrete counterexample beats a paragraph of doubt. Normal prose, no preamble, no padding, no self-assigned severity ratings. If it holds, say so in a line and say why.
Concrete examples of this angle in action
- Frontend: A component-ownership rule says "a feature module may import from shared, but not from another feature." Find the legitimate cross-feature case it forbids, or the back-channel (shared barrel re-export) that defeats the rule while satisfying it.
- Backend: A schema models
userandaccountas a clean one-to-one. Construct the real-world case (shared family account, service account, merged users) where one-to-one is violated and the constraint will throw in production. - DevOps: A k8s label taxonomy claims
tierandcomponentare independent axes. Find the workload that is both "tier=data" and "component=cache" ambiguously, so the same pod matches two mutually-exclusive selectors. - Security: A permission model defines roles as a strict hierarchy (admin ⊃ editor ⊃ viewer). Find the permission a viewer needs that an editor shouldn't have, breaking the subset assumption.
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.
- 2d ago First seen · 25 lines · 0 tokens per session scan A 687c6f6d17c1
formalist is an agent published in the GitHub repository Rebel028/gauntlet (5 stars, last pushed 1mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 471 tokens. 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
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.