FStarDev

A coding agent for the F* programming language compiler. It handles compiler source changes, the multi-stage bootstrap build, incremental rebuilding, and testing across compiler subsystems.

In plain words
What is it for?
Use it for F* compiler engineering, including parser, typechecker, SMT encoding, tactics, extraction, build-system, bootstrap, and snapshot tasks.
Why use it?
It helps developers work through a compiler codebase where changes may affect parsing, type checking, code generation, build stages, or the Pulse checker.

Agent

Install

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.

agentmods
npx agentmods add agents/fstarlang/fstar/fstardev
Clone the repo
git clone --depth 1 https://github.com/FStarLang/FStar
Per session 19 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 5,140 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00019 $0.05140
Opus 5 $0.00010 $0.02570
Sonnet 5 $0.00004 $0.01028
Haiku 4.5 $0.00002 $0.00514

Measured 3d ago against content hash a60bf3e09747, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

FStarDev 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 3d 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.

.github/agents/FStarDev.md · 460 lines

How it starts

The opening of the file, as written. The whole thing — 460 lines — stays where its author put it; the contents beside it link to each section on GitHub.

FStarDev Agent

Agent Identity

An expert F* compiler developer. This agent modifies the F* compiler sources, navigates the multi-stage bootstrap build, and validates changes through incremental rebuilds and testing. It covers compiler internals (parser, typechecker, extraction, SMT encoding, tactics), the Pulse checker plugin (stage 3), the build system (Make + dune), and the bootstrap/snapshot workflow.

Repository Layout

FStar/
├── src/                    # F* compiler sources (written in F*)
│   ├── basic/              # Core utilities, options, errors, platform
│   ├── syntax/             # AST definitions, free variables, substitution
│   ├── parser/             # Surface syntax AST, dependency analysis
│   ├── tosyntax/           # Desugaring: surface AST → core AST
│   ├── typechecker/        # Type inference, checking, normalization, NBE
│   ├── smtencoding/        # Encoding to SMT (Z3) queries
│   ├── extraction/         # Code generation: OCaml, F#, krml (C via KaRaMeL)
│   ├── tactics/            # Tactic engine, metaprogramming primitives
│   ├── reflection/         # Reflection API builtins
│   ├── interactive/        # IDE protocol, incremental checking
│   ├── prettyprint/        # A pretty-printer library (François Pottier's pprint)
│   ├── fstar/              # Top-level driver (FStarC.Main, FStarC.Universal)
│   ├── tests/              # Compiler unit tests (OCaml)
│   ├── ml/                 # Hand-written OCaml: lexer, parser (menhir), system utils
│   ├── class/              # Typeclasses used in the compiler
│   ├── data/               # Data structures used in the compiler
├── ulib/                   # F* standard library (verified .fst/.fsti files)
├── stage0/                 # OCaml snapshot: starting point for bootstrapped build.
├── stage1/                 # Built from: stage0 checks+extracts src/ → OCaml → dune build
├── stage2/                 # Built from: stage1 checks+extracts src/ → OCaml → dune build
├── stage3/                 # Stage2 + Pulse plugin linked in
├── pulse/                  # Pulse language implementation
│   ├── src/checker/        # Pulse typechecker (F* source, extracted to OCaml plugin)
│   ├── src/extraction/     # Pulse-specific extraction (to C, OCaml)
│   ├── src/syntax_extension/ # Pulse syntax desugaring
│   ├── src/ml/             # Hand-written OCaml for Pulse (parser .mly)
│   ├── lib/                # Pulse standard library (common/, pulse/)
│   ├── mk/                 # Build infrastructure (test.mk, common.mk, locate.mk)
│   ├── test/               # Pulse tests
│   └── share/pulse/examples/ # Pulse examples
├── karamel/                # KaRaMeL submodule (F*/Pulse → C extraction)
├── mk/                     # Top-level build infrastructure makefiles
├── tests/                  # F* integration tests
├── examples/               # F* examples
└── doc/                    # Documentation, book

Read the full file on GitHub · 460 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. 3d ago First seen · 460 lines · 19 tokens per session scan A a60bf3e09747

Subscribe to this mod's changes

FStarDev is an agent published in the GitHub repository FStarLang/FStar (3,101 stars, last pushed today), licensed Apache-2.0. It adds 19 tokens to every session and 5,140 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 2026-08-30.

Related

Other agents, from other repositories

proof-workflows

Agent "proof-workflows" from Verified-zkEVM/VCVio, covering proof workflows, proof strategy decision tree, monadic normalization with monadnorm, game-hopping recipe and step 1: state the security theorem.

Verified-zkEVM/VCVio · 0 tokens

interop

Interop/ is an experimental sibling to LatticeCrypto/ whose job is to let VCVio reason about Lean code emitted by Rust verification frontends.

Verified-zkEVM/VCVio · 0 tokens

p1-research-orchestrator

Phase 1 research pipeline orchestrator. Manages spec refinement via AskUserQuestion, exhaustive solution tree exploration with maximum parallel agents, sub-domain expert coordination, 3-round chief review, and structured artifact generation.

babyworm/rtl-agent-team · 50 tokens

p1-research-team-orchestrator

Phase 1 research team coordination teammate. Coordinates tree-of-thought solution exploration with parallel candidate deep-dive, sub-domain expert coordination, and 3-round chief review via TaskCreate/TaskList/TaskUpdate/SendMessage.

babyworm/rtl-agent-team · 55 tokens

p2-arch-team-orchestrator

Phase 2 architecture team coordination teammate. Coordinates dual-stream architecture design + C reference model development, dynamic convergence-based iterative review with wonder tracking via TaskCreate/TaskList/TaskUpdate/SendMessage.

babyworm/rtl-agent-team · 49 tokens

p4-implement-team-orchestrator

Phase 4 RTL implementation team coordination teammate. Coordinates 10-wave pipeline with per-module parallelism and inter-wave dependency graphs via TaskCreate/TaskList/TaskUpdate/SendMessage.

babyworm/rtl-agent-team · 46 tokens