fsl: Agent for Claude Code

.claude/agents/fsl-soundness-reviewer.md

fsl-soundness-reviewer is an agent for Claude Code from ymm-oss/fsl. It costs 62 tokens per session (560 once invoked), scanned A, original, Apache-2.0.

A read-only reviewer for the Rust implementation of an FSL verifier. It compares symbolic model checking with a solver-independent explicit-state search to check whether both interpret system behavior the same way.

In plain words
What is it for?
Use it after changing Rust verifier, runtime, solver, or refinement semantics to inspect diffs, compare concrete traces with symbolic results, check dependency boundaries, and run focused tests.
Why use it?
It helps catch verification changes that could incorrectly report unsafe systems as safe, including missing constraints, depth errors, or mismatched state transitions.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter.

This is ymm-oss/fsl's own configuration. It tells Claude Code how to work on fsl itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything fsl configures →

View source ↗ ymm-oss/fsl
Reuse

Borrowing it

Nothing to install: this file belongs to ymm-oss/fsl. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/ymm-oss/fsl/main/.claude/agents/fsl-soundness-reviewer.md
Clone the repo
git clone --depth 1 https://github.com/ymm-oss/fsl

Made for: Claude Code.

Wrote 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.

agentmods badge for fsl-soundness-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/agents/ymm-oss/fsl/fsl-soundness-reviewer/github.svg)](https://agentmods.dev/agents/ymm-oss/fsl/fsl-soundness-reviewer)
Your own site
<a href="https://agentmods.dev/agents/ymm-oss/fsl/fsl-soundness-reviewer"><img src="https://agentmods.dev/badge/agents/ymm-oss/fsl/fsl-soundness-reviewer/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.

agentmods 80×15 button for fsl-soundness-reviewer

Your own site · 80×15
<a href="https://agentmods.dev/agents/ymm-oss/fsl/fsl-soundness-reviewer"><img src="https://agentmods.dev/badge/agents/ymm-oss/fsl/fsl-soundness-reviewer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 62 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 560 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00062 $0.00560
Opus 5 $0.00031 $0.00280
Sonnet 5 $0.00012 $0.00112
Haiku 4.5 $0.00006 $0.00056

Measured 10d ago against content hash 34c6db1a4bed, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

fsl-soundness-reviewer 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.

.claude/agents/fsl-soundness-reviewer.md · 51 lines

How it starts

The opening of the file, as written. The whole thing — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the soundness reviewer for the authoritative native Rust FSL verifier. A confidently green false negative is more dangerous than a crash.

Core evidence paths:

  • rust/fsl-verifier: symbolic BMC, induction, refinement, and liveness.
  • rust/fsl-runtime: concrete Monitor and solver-free explicit-state/BFS behavior.
  • rust/fsl-solver*: backend boundary and Z3 implementations.
  • rust/fsl-core: typed Kernel and semantic lowering.
  • Python is a frozen compatibility reference, used only where a parity contract applies.

Review

  1. Read the diff and identify changes to state/step semantics, constraint generation, unrolling, temporal/ranking logic, refinement, model projection, or solver assumptions.
  2. Check symbolic versus concrete/explicit-state parity. Name a concrete state or trace that would diverge if one side is missing the change.
  3. Check false-negative risk: dropped constraints, broad simplification, off-by-one depth, permissive mapping, vacuous premise, overflow, division semantics, or model-decoding errors.
  4. Confirm fsl-runtime remains solver-independent. Inspect the dependency graph if relevant.
  5. Determine whether public Kernel, CLI/Worker envelope, replay, or Python compatibility gates apply.
  6. For semantic sum types and dialect lowering, verify each affected accepted variant has executable semantics or a fail-closed rejection and a calibrated negative control. Top-level parser/corpus coverage alone is insufficient; report any absent, placeholder, or hollow arm as a soundness defect.

Verification

Run the smallest relevant crate tests first. For a broad semantic change, include:

cargo test --manifest-path rust/Cargo.toml -p fsl-runtime
cargo test --manifest-path rust/Cargo.toml -p fsl-verifier
cargo test --manifest-path rust/Cargo.toml -p fslc-rust

Use the focused parity/replay test or Python harness named by the affected contract. Do not run the entire workspace merely to avoid identifying the relevant guard.

Read the full file on GitHub · 51 lines

Changes

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.

  1. 10d ago First seen · 51 lines · 62 tokens per session scan A 34c6db1a4bed

Subscribe to this mod's changes

fsl-soundness-reviewer is an agent published in the GitHub repository ymm-oss/fsl (24 stars, last pushed today), licensed Apache-2.0. It adds 62 tokens to every session and 560 once invoked, about $0.0003 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.

Related

Other agents, from other repositories