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.
npx agentmods add rules/mannansaood/termi/10-rust-coregit clone --depth 1 https://github.com/MannanSaood/termiWhat 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 | $0.00000 | $0.01099 |
| Opus 5 | $0.00000 | $0.00549 |
| Sonnet 5 | $0.00000 | $0.00220 |
| Haiku 4.5 | $0.00000 | $0.00110 |
Grade A, and why
10-rust-core 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 yesterday.
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 — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Rust Core Conventions
Build/test commands — use these exact invocations
# Fast loop — no Android SDK/NDK required, this is what you should run
# after almost every change:
cd rust && cargo test
# Before committing, also verify the JNI-boundary code still type-checks
# (it won't *run* here, but a signature drift should be caught):
cd rust && cargo check --features android --all-targets
# Host `--features android` still uses *host* libc. ioctl request types
# and missing constants (TIOCGPTN) only fail on the Android target.
# Needs `rustup target add aarch64-linux-android`. No NDK required.
cd rust && cargo check --target aarch64-linux-android --features android --lib
# Full lint pass matching CI exactly:
cd rust && cargo clippy --all-targets -- -D warnings -D clippy::unwrap_used -D clippy::expect_used
cd rust && cargo clippy --features android -- -D warnings -D clippy::unwrap_used -D clippy::expect_used
If cargo test fails to build because it's trying to pull in the jni
crate, check that you haven't accidentally added jni::-typed code (e.g.
jni::sys::jlong) to a function outside a #[cfg(feature = "android")]
block — see jni_safe.rs for the pattern of keeping handle-management
logic host-testable while gating actual JNIEnv calls.
Module map
vfs/capabilities.rs— the capability model: what operations (chmod/symlink/rename/...) are supported per filesystem type, with human-readable reasons. Pure logic, no I/O.vfs/provider.rs— theFsProvidertrait every storage backend implements.chmod/symlink/readlinkhave default "not supported" implementations — only override them if the backend genuinely supports the operation (seeInternalProviderfor the real Unix implementation).vfs/service.rs— the load-bearing piece.VfsServicechecks capabilities before dispatching to a provider and returns aVfsOutcome<T>(Ok/Blocked{reason,hint}/Degraded{value,caveat}) instead of a bareResult. Any new VFS operation should go through here, not directly through a provider, or the capability system silently stops applying to it.vfs/health.rs— SAF permission health-check state machine (PermissionState,PermissionProbetrait,HealthMonitor). The actual Android-side probe isn't wired up yet — see the TODO in that file and inandroid/.../PermissionManager.kt.jni_safe.rs— MANDATORY safety layer for all JNI calls. Never call rawJNIEnvmethods outside this module's wrappers. Split into an always-compiled pure half and anandroid-feature-gated half — see the module doc comment for why.utils/sync_ext.rs—LockExt::lock_safe(): use this instead of.lock().unwrap()on anyMutex. See module doc for when NOT to use it (mutexes protecting real invariants, not just data buffers).
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.
- yesterday First seen · 88 lines · 0 tokens per session scan A bd1b13197fd6
10-rust-core is a cursor rule published in the GitHub repository MannanSaood/termi (8 stars, last pushed yesterday), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,099 tokens. 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 cursor rules, from other repositories
cursorrules
You are an expert in SceneView, the #1 open-source 3D & AR SDK for Android and iOS.
compose-accessibility
Compose accessibility rules for TalkBack — icons, headings, Canvas, live regions, focus, modals, and attribute preservation.
compose-coroutines
Compose scroll/coroutine patterns — prevent programmatic scroll from being mistaken for user scroll, cancellation-safe flag resets.
ios-viewmodel
ViewModels conform to ObservableObject and expose state via @Published properties.
compose-ui
Compose UI conventions — Material 3, recomposition optimization, Compose-only UI, theming.
android-viewmodel
Android ViewModel conventions — StateFlow, repository abstraction, coroutines, no LiveData.