fix(memory): default null target to memory store

[teknium1]

Summary

memory(action="add", target=null, ...) now behaves like an omitted target and writes to the default memory store.

Source: nearai/ironclaw#4547 fixed the same strict-provider shape for IronClaw's memory_write capability. Hermes adapts the invariant at the Python tool dispatcher boundary instead of adding provider-specific schema plumbing.

Changes

• tools/memory_tool.py: normalize target is None to "memory" before target validation.
• tests/tools/test_memory_tool.py: add regression coverage for target=None and keep non-string targets rejected.

Validation

Check	Result
scripts/run_tests.sh tests/tools/test_memory_tool.py	71/71 passed
E2E	real MemoryStore, temp memory dir, target=None writes MEMORY.md; target=42 still rejected

Source adaptation

• IronClaw source PR: [codex] Fix memory_write null target handling nearai/ironclaw#4547
• Architectural difference: IronClaw strips optional null sentinels in first-party capability normalization; Hermes can fix this narrowly because the memory tool already has a documented default target.

Infographic

[github-actions]

🔎 Lint report: ironclaw-port/memory-null-target-4547 vs origin/main

ruff

Total: 0 on HEAD, 0 on base (➖ 0)

🆕 New issues: none

✅ Fixed issues: none

Unchanged: 0 pre-existing issues carried over.

ty (type checker)

Total: 10904 on HEAD, 10904 on base (➖ 0)

🆕 New issues (3):

Rule	Count
invalid-argument-type	2
invalid-assignment	1

First entries

tests/tools/test_memory_tool.py:429: [invalid-argument-type] invalid-argument-type: Argument to function `memory_tool` is incorrect: Expected `str`, found `None`
tests/tools/test_memory_tool.py:440: [invalid-argument-type] invalid-argument-type: Argument to function `memory_tool` is incorrect: Expected `str`, found `Literal[42]`
tests/run_agent/test_credits_notices_toggle.py:76: [invalid-assignment] invalid-assignment: Object of type `None` is not assignable to attribute `_credits_session_start_micros` of type `int`

✅ Fixed issues (2):

Rule	Count
unresolved-attribute	2

First entries

run_agent.py:2920: [unresolved-attribute] unresolved-attribute: Object of type `Self@get_credits_spent_micros` has no attribute `_credits_session_start_micros`
tests/run_agent/test_credits_notices_toggle.py:76: [unresolved-attribute] unresolved-attribute: Unresolved attribute `_credits_session_start_micros` on type `AIAgent`

Unchanged: 5733 pre-existing issues carried over.

Diagnostics are surfaced as warnings — this check never fails the build.

[liuhao1024]

Verification: Reviewed the diff — treating target=None (JSON null from strict providers) as the default "memory" store is a clean defensive fix. The coercion happens before the target not in {"memory", "user"} guard, so null values won't leak through. Test coverage includes null-target default and non-string rejection. CI green.

[teknium1]

Merged in PR #46356 — memory(target=null) now falls back to the default memory store instead of failing validation.