
Migrate Legacy Code to Safe Rust
Guides an agent through porting legacy C++, Fortran, Java, Python or C code to safe, optimized Rust with telemetry.
- Repository
- google/skills
- Stars
- 21,782
- Licence
- Apache-2.0
- Files
- 3 files
What it does
Once installed, the agent follows a staged playbook to take an existing codebase from a GitHub URL or local folder and port it to Rust. It first records baseline measurements of the legacy code, then writes an architecture plan, implements a Rust version with tests and output-parity checks, and optionally applies profile-guided optimization. It logs timings and results to a JSON file via a bundled Python script and can produce a written report and an HTML summary page.
Great for
- Engineers planning or running a migration of a legacy C++, Fortran, Java, Python or C project to Rust
- Teams who want measured before-and-after benchmarks and a documented parity check against the original program
Skip it if
- Quick one-off Rust syntax questions or small unrelated refactors, which the skill itself says it should not be used for
- Running untrusted cloned repositories without review: the skill executes cloned build and run commands after confirmation, which can run arbitrary code from that repo
Try asking
- “Migrate this C++ repo at github.com/example/solver to safe Rust and benchmark it against the original”
- “Port the Fortran code in ./physics_sim to Rust and check output parity across 10 random seeds”
- “Audit my legacy Python data pipeline and produce a Rust architecture blueprint”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 3 original files, the Apache-2.0 licence and an attribution note.
migrate-to-rust.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip migrate-to-rust.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip migrate-to-rust.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Review the build and run commands the agent proposes in Stage 0 before confirming, since they will execute code from the cloned repository.
- Point the agent at a local copy you have already inspected when the source is untrusted.
- Expect the workflow to take a long time; it caps self-healing at 5 iterations per loop and reports remaining gaps rather than looping forever.
What's inside
- references/pgo_and_telemetry_guide.md
- scripts/telemetry.py
- SKILL.md
Redistributed unmodified from google/skills under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.