LlmCore.Agent.Components.ParseToolCalls (llm_core v0.5.0)

Copy Markdown View Source

Extracts tool calls from the LLM response.

If the response contains no tool calls (or an empty list), sets decision to {:done, response} — the LLM has produced a final text response and no further iteration is needed.

If tool calls are present, populates ctx.tool_calls for downstream stages.

When ctx.terminal_tool is set and a matching call is present, the pipeline marks the response as done and stores the matching call and raw arguments on the context. The call is not validated, dispatched, or injected into the next turn.

Analogous to a context merge stage: takes raw input and normalizes it into the pipeline's working format.

Summary

Functions

call(ctx, opts)

Extracts tool calls from ctx.response.tool_calls.

Functions

call(ctx, opts)

@spec call(
  LlmCore.Agent.Context.t(),
  keyword()
) :: LlmCore.Agent.Context.t()

Extracts tool calls from ctx.response.tool_calls.

Short-circuits when ctx.status is :error.

Parameters

  • ctx — %Context{} with the LLM response
  • opts — ALF stage options (unused)

Returns

Updated %Context{} with either:

  • tool_calls populated and pipeline continues, or
  • decision: {:done, response} when no tool calls are present
  • decision: {:done, response} plus terminal fields when the terminal tool is called