Ainova SystemsAinova Systems
Open source · AI intelligence CLI

Intelligence

Maintain AI instructions once, then render the native files each enabled tool expects. Intelligence versions the source, packages and resolved state through one manifest, lockfile and CLI.

One Source, Native Outputs

AI tools move quickly. The project keeps its instructions and package state reviewable across those changes.

Source
Project-owned rules, agents and skills stay in tool-neutral Markdown under intelligence/.
Resolution
intelligence.yaml records intent. The committed intelligence.lock records the exact package source and revision.
Rendering
Adapters translate the same source into the native rule, agent and skill formats of each enabled AI tool.
Recovery
A fresh clone restores missing packages directly from the lockfile before rendering, preserving the exact package state.

Controls You Can Inspect

Tool coverage

Six AI Tools from One Source

Claude Code, Cursor, GitHub Copilot, Codex, Pi and OpenCode receive the formats and scoping their own surfaces support.

  • Scoped rules where supported
  • Native agent formats
  • Complete skill bundles
Change control

Preview Changes Before Writing

Initialization, conversion and updates expose a preview path. Project alignment remains explicit in CI and tracked state stays reviewable.

  • Preview modes
  • Explicit application
  • CI consistency checks
Sync boundary

Restore Outputs When Sync Fails

Managed output paths are snapshotted before sync. A later adapter failure restores earlier outputs to their pre-run state.

  • Ownership validation
  • Run-level snapshots
  • Failure restoration

Inspect the Intelligence CLI

The GitHub repository carries the source, documentation, changelog and current release state.