Project records

These are historical design records: plans and research written while features were being designed. Most pages state their status and the date or commit they were reviewed against, but none is maintained as current documentation. For how Praetor behaves today, read the Guides, the HISS specification and the architecture decisions.

Plans

• Bidirectional module contracts: audited design proposal for contracts between generated configuration, observations and repairs.
• Fleet drift and schema conformance: design proposal for declared copies and cross-repository payload schema checks, with the read-only harvest drift survey that ships first.
• IDE-driven agent setup: the goal of IDE extensions that set up and reconcile every agent effective policy selects.
• Infrastructure development environments: researched strategy for VM and cluster backends, not implemented.
• Shared lifecycle management: staged plan for agent session, workspace and state lifecycles.
• Package development pipeline: proposed configurable package development and reuse pipeline.
• Shared planning and reconciliation: incremental plan from the planning draft compiler to graph reconciliation.
• Upstream contribution workflow: proposed upstream contributions within the package pipeline.
• Wish platform integration: plan for forge and Discord projections of the wishes ledger.

Research

• Compact management data: locally benchmarked research on compact, configurable management data.
• Public dogfood capability discovery: the discovery run over 100 public repositories and what it found.
• Go package reuse: which mature Go packages to reuse instead of maintaining custom code.
• HISS refinement audit: audit and scoped repair of the HISS rules.
• IDE agent bootstrap: VS Code agent bootstrap and safe praetorctl execution.
• Infrastructure sandboxes: isolation tiers for template-driven development.
• Wish voting platforms: ballots and votes across forges and Discord.