This adds experimental support for virtual models. A virtual model is a catalog entry that an extension registers with pi.registerVirtualModel(). It does not talk to a provider itself; for every request it picks a physical model and thinking level. The motivation is to make routing policies pluggable: plan on a strong model and implement on a cheap one, pick a model with a classifier, or move to a larger window when the context grows. Today this needs model switching by hand or hacks in prepareRequest.
The selection stays virtual: model_change records it, resume restores it, and each response records the physical model and thinking level that produced it. Router state can be stored on the session branch, so it follows /tree and forks and survives compaction. Context limits, compaction and image sizing use the physical model a request was routed to. examples/extensions/jev-router.ts shows a full router that uses the Jev classifier to pick Sol or Terra for planning and hands off to Luna after the first edit. docs/virtual-models.md describes the API.
Replace the tsgo native preview with typescript@7.0.2 and port the check scripts to the TS 7 API. Target ES2024, drop useDefineForClassFields: false and unused decorator options, and enable verbatimModuleSyntax. Replace tsx with node plus the source resolver hook, passed to --import as a file URL.
closes#9965
This change makes system prompt text and tool changes part of the transcript rather than silently rewriting its starting conditions. This lets Pi record when instructions changed or tools became available, restore that state after resuming or navigating branches, and preserve cached prompt prefixes where the upstream supports it.
`parseFrontmatter` runs a real YAML parser, so `tools: [read, bash]`
arrives as an array while `tools: read, bash` arrives as a string. The
loader only handled the string form and called `.split(",")`, which
throws on the array form.
The throw escapes `loadAgentsFromDir` -- only `readdirSync` and
`readFileSync` are guarded -- so one agent file using the array form
broke discovery for every agent in that scope, not just itself.
Normalize both spellings in one helper, and type the frontmatter fields
as `unknown` so the string assumptions are explicit. The previous
`parseFrontmatter<Record<string, string>>` annotation asserted a shape
the YAML parser never guarantees, which is what hid the mismatch from
tsc while `AgentConfig.tools` was already typed `string[]`.
Move provider auth and OAuth flows onto pi-ai Models, compose models.json and extension overlays through ModelRuntime, and retain ModelRegistry as an extension compatibility facade.
Armin Ronacher