executeSingleTool had three exit paths but only the success path
emitted all end signals. The interrupted path (nested agent
approval) skipped OnToolEnd and StreamEventToolEnd entirely,
leaving hook consumers with an unpaired OnToolStart. The error
path also missed StreamEventToolEnd and AgentHooks.OnToolEnd.
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
When an MCP server sets StructuredContent without populating
Content with TextContent entries, extractMCPContent returned
an empty string, making successful tool calls look empty to
the agent. Now the function serializes StructuredContent as
JSON when no text parts are found.
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
buildTools was only copying "properties" and "required" into
ToolInputSchemaParam, silently dropping root-level constraints
like additionalProperties and $defs. Forward them through the
SDK's ExtraFields map so the full schema reaches the API.
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
reflect.TypeOf on a nil interface value returns nil, causing a
panic when Kind() is called. Use reflect.TypeFor[T]() instead,
which resolves the type directly from the type parameter.
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
Fix Wait draining events from concurrent consumers
Wait() was ranging over the public Events channel, competing
with any concurrent reader for events. Callers that streamed
events in one goroutine and called Wait() in another would
lose an arbitrary subset of events. Wait now only blocks on
the done channel; the result fields are already visible thanks
to the close ordering (set fields → close events → close done).
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
StreamAccumulator never set its model field, so
Response().Model was always empty for streamed
completions. Add a Model field to stream events
and populate it in all three providers (OpenAI,
Anthropic, Bedrock).
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
FrequencyPenalty, PresencePenalty, and ParallelToolCalls were
present on ChatCompletionRequest but never forwarded in
buildParams, so agent-level settings configured via functional
options were silently ignored for the OpenAI provider.
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
buildResponseFormat was hardcoding strict: true instead of
reading rf.JSONSchema.Strict, making it impossible to request
non-strict JSON-schema output.
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
jsonSchemaFor panicked on unsupported types, which meant
FunctionTool, NewOutputType, and RunTyped would crash the
process during setup rather than returning a normal error.
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
The agentTool.Execute method accepted {} despite the schema
marking input as required. Unlike functionTool, it skipped
required-field validation, silently sending an empty message
to the sub-agent.
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
When the LLM returns tool calls after a handoff in the same
assistant message, they were silently dropped. This left
orphaned tool_call entries without matching tool-result
messages, causing protocol errors on the next LLM turn.
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
Incrementing toolIndex on every ContentBlockStop caused it to drift
when non-tool blocks (e.g. text) preceded tool-use blocks, making
Bedrock tool calls disappear from the accumulated response.
Track whether the current content block is a tool-use block and only
bump the index when a tool-use block ends.
Signed-off-by: Bryan Frimin <bryan@getprobo.com>
- Fix redirection after document deletion in detail page
- Fix refetch after publication
- Return proper error when trying to update a published document
Signed-off-by: Sacha Al Himdani <sacha@getprobo.com>
