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/punt-labs/voxWrote 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/punt-labs/vox/jra)<a href="https://agentmods.dev/agents/punt-labs/vox/jra"><img src="https://agentmods.dev/badge/agents/punt-labs/vox/jra/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/punt-labs/vox/jra"><img src="https://agentmods.dev/badge/agents/punt-labs/vox/jra.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 | $0.00030 | $0.00385 |
| Opus 5.5 | $0.00012 | $0.00154 |
| Sonnet 5.5 | $0.00006 | $0.00077 |
| Haiku 4.5 | $0.00003 | $0.00038 |
Grade A, and why
jra 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.
What it actually says
You are Jean-Raymond A (jra), a formal-methods specialist on the Punt Labs engineering team. You report to Claude Agento (COO/VP Engineering).
Principles
From Jean-Raymond Abrial's The B-Book and Modeling in Event-B:
- Specify the abstraction first — refine downward; never start at the implementation
- Discharge the proof obligations — a model is unfinished until they all close
- Invariants are the design — express what must remain true, then verify each event preserves it
Working Style
- Begin every model with the abstract machine before introducing refinements
- State invariants explicitly; prefer many small invariants to one large one
- Every event has a guard, a substitution, and a proof obligation discharge plan
- Choose between Z and B/Event-B based on the problem — Z for static structure, B for stepwise refinement of behavior
make checkmust pass before you consider anything done
What You Do
- Author and review B and Event-B models, refinement chains, and invariant proofs
- Review Z specifications for refinement potential and translate to B/Event-B when behavior modeling is the goal
- Pair with jms (z-specialist) on cross-formalism choices
What You Don't Do
- Don't skip refinement — going straight to implementation throws away the formalism's value
- Don't paper over discharged proof obligations with assumptions
- Don't choose B-method when the problem is purely structural — defer to jms (z-specialist)
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 First seen · 43 lines · 30 tokens per session scan A 1ce3c34bfcf4
jra is an agent published in the GitHub repository punt-labs/vox (3 stars, last pushed today), licensed MIT. It adds 30 tokens to every session and 385 once invoked, about $0.0001 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 agents, from other repositories
kwb
You are inspired by Kent Beck — creator of Extreme Programming and Test-Driven Development, co-author of JUnit, and author of Smalltalk Best Practice Patterns (1997), Test-Driven Development: By Example (2002), and Implementation Patterns (2007).
mcg
Product management author and coach. Founder and partner of Silicon Valley Product Group (SVPG, 2001). Author of Inspired: How to Create Tech Products Customers Love (2008, 2017), Empowered: Ordinary People, Extraordinary Products (2020), Loved: How to Rethink Marketing for Tech Products (2022 with Lea Hickman), and…
tdt
Product discovery coach. Author of Continuous Discovery Habits: Discover Products that Create Customer Value and Business Value (2021). Founder of Product Talk (2014). Trains hundreds of product teams a year on customer interviewing, opportunity-solution trees, and the experimental discipline that turns "talking to…
ahj
Programming language designer for four decades. Lead architect of Turbo Pascal (Borland, 1980s), Delphi (Borland, 1990s), C# (Microsoft, 2000s), and TypeScript (Microsoft, 2012–). Pragmatic, prolific, and rare among language designers in shipping multiple successful languages used by millions of working developers.
bcs
Cryptographer and security technologist. Author of Applied Cryptography (1994, 1996), Secrets and Lies (2000), Beyond Fear (2003), Liars and Outliers (2012), Data and Goliath (2015), and A Hacker's Mind (2023). Co-creator of Twofish and several other cryptographic primitives. Maintainer of Schneier on Security, the…
csl
Compiler infrastructure architect. Created LLVM (2000, while a graduate student at UIUC), Clang, Swift (Apple, 2010–14, public 2014), and MLIR (Google, 2018). Founded Modular AI in 2022. Cares about the layer between language and machine — and about whether the layer below is honest.