unified-core-runtime
Skill177 starsChanged 7 days agoApache-2.0View on GitHub
Assemble source-built modules into one Cyber runtime with shared tools and configuration; validate without operating devices.
In this repository19 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .github/copilot-instructions.mdCopilot instructions
- agent-evaluationSkill
- algorithm-developmentSkill
- buildSkill
- change-reviewSkill
- container-buildSkill
- engineering-solution-designSkill
- interface-migrationSkill
- issue-localizationSkill
- lintSkill
- long-horizon-developmentSkill
- model-artifact-integrationSkill
- multi-repo-devSkill
- performance-optimizationSkill
- repository-agent-setupSkill
- simulation-regressionSkill
- test-impact-analysisSkill
- vendorSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 15
All processes use one authoritative Cyber configuration root and explicitly coordinated domain/network/clock settings.
- Distinctive: line 47
Keep source, archive and installed layouts equivalent and relocatable.
- Distinctive: line 58
No physical CAN, camera or LiDAR is initialized during automated validation.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Unified core runtime implementation
Agreed principles
- Bzlmod manages source dependencies. One final integration root resolves core, messages, tools, hardware and algorithm modules with the same toolchain.
- Install one core tool set: mainboard, cyber_launch, cyber_recorder and cyber_monitor. Modules contribute dynamically loaded components, not another middleware/tool installation.
- All processes use one authoritative Cyber configuration root and explicitly coordinated domain/network/clock settings. Process-specific scheduler files may differ within that root.
- Device, vehicle and algorithm configuration belongs to its module; it must not replace the common middleware configuration.
- Repository boundaries do not force process boundaries. Choose one or several mainboards according to execution and failure isolation requirements.
- Full-source integration is the current scope. Do not introduce binary ABI release frameworks, separate per-module Cyber roots, or standalone/managed delivery modes as prerequisites.
- Install declared files and dependency runfiles; never infer dependencies by scanning host libraries or concatenating separate product environments.
- Module selection/configuration must be explicit. Installation never starts devices. Recorder/monitor retain caller-relative data paths; recordings and logs are outside the installation.
Implementation procedure
- Read the
Core runtime contract, public
pkgs/core/runtimerules, hardware compatibility exports, module targets and relevant configuration. - Declare module contributions with core's
runtime_module/RuntimeModuleInfo: components, explicit component/runtime resource paths, dependency metadata, business configuration, DAG/launch and resources. Infra repositories export theirproducts/*:moduletargets. - Compose those contributions in a final Bzlmod integration root, such as
tools/system, adding core tools/config exactly once. Reject module-owned middleware/tool destinations and conflicts. The common runtime namespace isconf/,dag/,launch/,share/,lib/components/, and core-ownedbin/; never derive deployment paths from repository names or Bazel output roots. - Reuse core's runtime layout, wrappers, runfiles and rules_pkg delivery. Keep source, archive and installed layouts equivalent and relocatable.
- Validate inside the managed container as the mapped non-root user. Use
--config=devfor approved source-checkout validation; do not rewrite the release lock or publish local paths. Unpublished APIs/modules require explicit source overrides; local success is not registry release validation. - Exercise FakeCAN and synthetic sensor modules together through one runtime: common configuration, pub/sub, one recorder capturing both schemas, monitor, independent shutdown, archive/install relocation and missing-input errors.
Acceptance and boundaries
No physical CAN, camera or LiDAR is initialized during automated validation. Synthetic fixtures are private/test-only and absent from production packages. Local build/runtime success is not remote publication, hardware acceptance or permission to actuate. Record failures and evidence outside this skill.
This skill is the requested design/implementation entry under .agents.
Long-term source-backed runtime details remain in the canonical context
document rather than a second knowledge tree.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

