fix: make ToolContext hashable to match RunContextWrapper

[ioleksiuk]

Summary

ToolContext (src/agents/tool_context.py) is decorated with bare @dataclass, which defaults to eq=True and therefore sets __hash__ = None, making instances unhashable. Its parent RunContextWrapper (src/agents/run_context.py) is decorated @dataclass(eq=False) and is hashable by identity. The subclass silently dropped the parent's hashability contract, so user code that puts a tool context into a set, dict key, WeakKeyDictionary, lru_cache, etc. — a reasonable thing to do given the parent supports it — fails with TypeError: unhashable type: 'ToolContext'.

Repro against main / v0.15.1:

from agents.run_context import RunContextWrapper
from agents.tool_context import ToolContext

hash(RunContextWrapper(context=None))                                            # OK
hash(ToolContext(context=None, tool_name="t", tool_call_id="c", tool_arguments="{}"))
# TypeError: unhashable type: 'ToolContext'

Fix: switch the decorator to @dataclass(eq=False) so the subclass preserves the parent's identity-based hashing. This is a one-line change with no behavior impact for code that does not rely on hashing — ToolContext already defines its own __init__, so the dataclass-generated __eq__ was the only thing the bare decorator was adding, and identity equality is what the parent intentionally uses.

Test plan

• Added tests/test_tool_context.py::test_tool_context_is_hashable_like_run_context_wrapper covering hash(ToolContext(...)), equality of repeated hashes, and use as a dict key.
• uv run pytest tests/test_tool_context.py -v — all 15 tests pass.
• make format, make lint — clean.
• make typecheck — no new errors in the touched files (pre-existing errors in voice/sandbox modules are unrelated).
• make tests — same 12 pre-existing failures as main (sandbox/runloop and voice modules); no new failures.

Issue number

N/A

Checks

• I've added new tests (if relevant)
• I've added/updated the relevant documentation
• I've run make lint and make format
• I've made sure tests pass

[seratch]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Another round soon, please!

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[ioleksiuk]

Thank you for the quick response, @seratch ! 🙌