feat(actor): add tool-calling runtime for execution actors #152

New issue

Closed

opened 2026-02-22 23:39:39 +00:00 by freemo · 0 comments

freemo commented

2026-02-22 23:39:39 +00:00

Owner

Metadata

Commit: feat(actor): add tool-calling runtime for execution actors
Branch: feature/m1-actor-runtime

Subtasks

Add tool-calling loop that maps ToolRegistry specs to provider tool schemas and executes through ToolRunner with max-iteration safeguards.
Capture tool call metadata (tool name, inputs, outputs, duration, success) into a structured record for downstream ChangeSet/decision usage.
Thread sandbox root + resource bindings into tool inputs (default to plan/project resources).
Update docs/reference/actor_runtime.md with the tool-call loop and error semantics.
Tests (Behave): Add scenarios for tool-call loop, error handling, and max-iteration termination.
Tests (Robot): Add Robot smoke test running agents actor run with a tool-calling actor against a temp repo.
Tests (ASV): Add benchmarks/actor_runtime_bench.py for tool-call loop overhead.
Verify coverage >=97% via nox -s coverage_report. If coverage is <97% then review the current unit test coverage report at build/coverage.xml and use it to write new Behave based unit tests to improve coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun nox -s coverage_report to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
Run nox (all default sessions, including benchmark), fix any errors if needed ensuring nox passes across entire code base, do not ignore any failure even if it seems unrelated to this commit, fix it.

Section: #### M1: Minimal Local Source-Code Workflow (Target: Day 7, recovery path)
Status: Completed

## Metadata - **Commit**: `feat(actor): add tool-calling runtime for execution actors` - **Branch**: `feature/m1-actor-runtime` ## Subtasks - [x] Add tool-calling loop that maps ToolRegistry specs to provider tool schemas and executes through ToolRunner with max-iteration safeguards. - [x] Capture tool call metadata (tool name, inputs, outputs, duration, success) into a structured record for downstream ChangeSet/decision usage. - [x] Thread sandbox root + resource bindings into tool inputs (default to plan/project resources). - [x] Update `docs/reference/actor_runtime.md` with the tool-call loop and error semantics. - [x] Tests (Behave): Add scenarios for tool-call loop, error handling, and max-iteration termination. - [x] Tests (Robot): Add Robot smoke test running `agents actor run` with a tool-calling actor against a temp repo. - [x] Tests (ASV): Add `benchmarks/actor_runtime_bench.py` for tool-call loop overhead. - [x] Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%. - [x] Run `nox` (all default sessions, including benchmark), fix any errors if needed ensuring nox passes across **entire** code base, do not ignore any failure even if it seems unrelated to this commit, fix it. **Section**: #### M1: Minimal Local Source-Code Workflow (Target: Day 7, recovery path) **Status**: Completed

Rows
Columns