Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/gertsylvest/meta-teamnpx agentmods add skills/gertsylvest/meta-team/miri-dspWrote 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/skills/gertsylvest/meta-team/miri-dsp)<a href="https://agentmods.dev/skills/gertsylvest/meta-team/miri-dsp"><img src="https://agentmods.dev/badge/skills/gertsylvest/meta-team/miri-dsp/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/skills/gertsylvest/meta-team/miri-dsp"><img src="https://agentmods.dev/badge/skills/gertsylvest/meta-team/miri-dsp.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 | $0.00073 | $0.00762 |
| Opus 5 | $0.00036 | $0.00381 |
| Sonnet 5 | $0.00015 | $0.00152 |
| Haiku 4.5 | $0.00007 | $0.00076 |
Grade A, and why
miri-dsp 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 11d 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 — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Miri DSP
miri interprets Rust's mid-level IR and catches a large class of undefined behaviour that cargo test will silently miss: out-of-bounds slice access, use-after-free, reads of uninitialised memory, invalid unsafe pointer arithmetic, and aliasing violations under the Stacked Borrows / Tree Borrows models.
Miri is most valuable on no_std DSP crates — they tend to contain hand-written unsafe for SIMD and ringbuffers where the compiler cannot prove safety. The cost is runtime speed (Miri is an interpreter, not a compiler), so this skill targets the unit tests of a single crate rather than a whole workspace.
Limitations
- Miri does not target
wasm32. It runs on the host triple. UB-free under Miri does not guarantee UB-free under wasm32 — but in practice the UB classes Miri catches are target-independent. - Floating-point intrinsics are slow under Miri. Heavy DSP benchmark suites will time out. Focus tests on correctness invariants and
unsafeboundaries, not on full audio renders. cargo benchdoes not run under Miri — use thecargo-bench-rtskill on native instead.
Requirements
- Rust nightly toolchain with the miri component:
rustup toolchain install nightly rustup +nightly component add miri
The script bootstraps both if missing.
Instructions
The argument is in $ARGUMENTS. Pass it directly to the run script:
bash "$(dirname "$0")/run.sh" $ARGUMENTS
The argument is the path to the crate directory.
Output sections
| Section | What it reports |
|---|---|
TOOLCHAIN |
Whether nightly + miri are installed |
MIRI |
Result of cargo +nightly miri test |
Typical usage
# Run miri on a single DSP crate
bash run.sh ./rust/audio-dsp
# Run on every DSP crate in a workspace
for crate in rust/dsp-*; do bash run.sh "$crate"; done
What to look for
MIRIfailing with "Undefined Behavior" — read the report carefully. Miri pinpoints the exact statement and explains the violation (e.g. "reading uninitialised bytes", "out-of-bounds pointer offset", "stacked borrows violation").MIRItiming out — narrow the test set. Mark long-running renders with#[cfg(not(miri))]so they are skipped under Miri while still running undercargo test.MIRIreporting "unsupported operation" — Miri does not support every libc or syscall. If the offending call is in a third-party dep, gate that path behind a feature flag the DSP crate does not enable; if it is in your own code, the operation probably should not be in a DSP crate at all.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 11d ago First seen · 66 lines · 73 tokens per session scan A 7b94f34bc4fc
miri-dsp is a skill published in the GitHub repository gertsylvest/meta-team (5 stars, last pushed 1mo ago), licensed MIT. It adds 73 tokens to every session and 762 once invoked, about $0.0004 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 2026-08-31.
Other skills, from other repositories
edge-python
Write, run, test and package Edge Python programs with the edge CLI. Use when editing .py files in an Edge Python project or when the user asks for Edge Python code.
cargo-fuzz
Sets up and runs cargo-fuzz, the standard fuzzing tool for Cargo-based Rust projects. Covers cargo fuzz init, the nightly toolchain requirement, fuzztarget! harnesses, Arbitrary-derived structured inputs, sanitizer options, cargo fuzz coverage, and reproducing a crash artifact. Use when fuzzing a Rust crate, writing a…
mutation-test
Run mutation tests with cargo-mutants to verify test quality. Reports caught, missed, and timed-out mutants.
check-impl
Implementation verification - runs format, lint, test, and build checks on the workspace.
rust-development
Idiomatisk Rust-utvikling med cargo, clippy, error handling, async/tokio, unsafe og testing.
cargo-fuzz
Use when initializing, running, measuring coverage, or triaging a cargo-fuzz target in a Rust crate. Not for remote, credential, publish, deploy, or irreversible changes.