| Age | Commit message (Collapse) | Author |
|---|
|
Teach provider config and auth resolution about the Codex Responses
dialect, including user-facing `style = "openai_responses"` with
`dialect = "codex"`. Serialize and parse Responses traffic with
provider-specific reasoning replay, assistant phase metadata, and robust
function-call assembly keyed by `output_index` so streamed tool inputs
survive proxy quirks and empty terminal payloads.
Also persist thinking origins and message metadata across sessions, add
the Anthropic interleaved-thinking header switch, write per-session
debug logs, and improve the TUI and scripts for inspecting tool output
and session costs.
|
|
Add the OpenAI Responses API wire dialect (openai_responses APIStyle +
OpenAIResponsesConfig), a request serializer (instructions/input items,
flat function tools, reasoning, store:false, include encrypted reasoning),
and a streaming state machine over the typed response.* SSE events mapping
to the shared block model. Wire dispatch, buildProviderConfig, the C-ABI
free switch, and the reasoning selectors. Ship a commented Codex example in
the default config.
Implemented from the Responses API docs + the open-source Codex client and
covered by fixture tests; live verification against a ChatGPT plan is still
required (documented).
|
|
Add provider.mergeHeaders and use it in both the openai_chat and
anthropic_messages providers so config extra_headers ride on the model
request, merged onto the built-in header set.
|
|
|
|
The tool set is no longer part of the per-turn Config snapshot. Agent now
owns a ToolRegistry (created empty at init), populated via the new
Agent.registerTool / registerToolSource methods. openStream and
OpenStreamFn take the registry as an explicit parameter rather than reading
it from cfg.registry, so swapping provider/model between turns (setConfig)
no longer disturbs the tool set.
CLI migrated to register the Lua tool source onto the agent instead of a
locally-owned registry. Test harnesses updated to stage tools on the agent.
|
|
converted the main agent loop from push-based (into callbacks on a
Receiver vtable) to pull-based, where `next()` re-enters a state-machine
Stream until the next event can be returned.
|
|
|
|
|
|
Replace all references to the "system" layer with "base" to better reflect
its role as the foundational extension/tool layer in panto's hierarchy.
The layer hierarchy now consistently uses: project > user > base.
Includes:
- Update agent README and build.zig documentation
- Refactor tool registry and config module to support layered lookups
- Add TestHarness abstraction for cleaner test setup
- Improve JSON serialization with wire-encoded tool names
- Add glob pattern matching for tool/extension discovery
|
|
|
|
OpenAI and Anthropic have tool details (id, name) converge on different
timelines. In all cases they are available in `onBlockComplete` which
comes with the full block, and previously we had an optional `BlockMeta`
argument to `onBlockStart` - the anthropic timeline.
We removed the `BlockMeta` from `onBlockStart` since it was always going
to be unreliable, and now have an explicit `onToolDetails` instead. This
allows us to provide the tool id and name as early as possible from the
openai provider (could be in the middle of content deltas), and provide
a reliable signal across providers.
Also, fix test output.
|
|
Close each tool_use on next-index delta (and at finalize) so receivers
see one start/delta/complete trio per tool, matching Anthropic. Log+drop
fragments for already-closed indices.
Remove BlockMeta from the Receiver interface: tool id/name can't be
reliably known at onBlockStart under OpenAI's streaming model. Receivers
get identity from the assembled ContentBlock at onBlockComplete. CLI
renders 'tool: {args} : (name)' accordingly.
|
|
|
|
|
|
|
|
|
