fix(file-safety): return POSIX inner_path from classify_sandbox_mirro…

[aturzone]

…r_target on Windows

classify_sandbox_mirror_target built inner_path with str(Path(*parts)), which emits OS-native separators — "profiles\group1\SOUL.md" on Windows. inner_path is a logical HERMES_HOME-relative identifier surfaced to the model ("the authoritative file is likely <inner_path>"), not an on-disk path, and its classify_container_mirror_target sibling already builds it with .as_posix(). Emit it POSIX-style so the two mirror classifiers agree and the hint stays canonical on every platform. mirror_root stays native (str) — it names a real host dir for display, matching the sibling.

tests: the sandbox-mirror suite already encodes the POSIX inner_path contract but only ran green on POSIX CI (the project's CI is Linux-only); on native Windows five cases failed. Normalize the two mirror_root assertions the same way the container-mirror tests already do, and add a regression guard that inner_path never contains a native separator.

Assisted-by: Claude Code (Opus 4.8)

What does this PR do?

Related Issue

Fixes #

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 🔒 Security fix
• 📝 Documentation update
• ✅ Tests (adding or improving test coverage)
• ♻️ Refactor (no behavior change)
• 🎯 New skill (bundled or hub)

Changes Made

How to Test

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/ -q and all tests pass
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform:

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A
• I've updated tool descriptions/schemas if I changed tool behavior — or N/A

For New Skills

• This skill is broadly useful to most users (if bundled) — see Contributing Guide
• SKILL.md follows the standard format (frontmatter, trigger conditions, steps, pitfalls)
• No external dependencies that aren't already available (prefer stdlib, curl, existing Hermes tools)
• I've tested the skill end-to-end: hermes --toolsets skills -q "Use the X skill to do Y"

Screenshots / Logs

[teknium1]

Thanks for the focused cross-platform fix. The premise holds on current main: classify_sandbox_mirror_target() builds the model-facing logical inner_path with str(Path(...)) at agent/file_safety.py:564, while its container-mirror sibling already uses inner.as_posix() at agent/file_safety.py:648. The warning exposes this value at agent/file_safety.py:595 through the file-tool guard (tools/file_tools.py:776).

The PR aligns those two classifiers and adds a direct separator regression assertion. GitHub reports the PR as mergeable against current main.

Automated hermes-sweeper review.

[monerostar]

Ubuntu 26.04 on linux-5800x.

Live call of classify_sandbox_mirror_target on main already returns POSIX inner_path here (profiles/group1/memories/MEMORY.md, no backslashes), which is expected on Linux. The real footgun is Windows str(Path(...)).

PR side:

• same live path still POSIX
• tests/agent/test_file_safety_sandbox_mirror.py: 14 passed (0.08s), including test_inner_path_uses_posix_separators
• aligns sandbox mirror hints with the container-mirror .as_posix() sibling

Looks good.