fix(context): count fallback compactions as ineffective

[LeonSGP43]

What does this PR do?

Treats deterministic fallback compactions as degraded attempts in the existing anti-thrash guard.

Today, a failed summary call can still shrink the prompt enough that the next real prompt_tokens reading falls below threshold and resets _ineffective_compression_count back to 0. That lets repeated empty-summary fallback rewrites loop indefinitely without ever tripping the existing breaker. This patch keeps the real-usage-based guard intact, but consumes a strike whenever the completed compaction had to fall back to the placeholder summary.

Related Issue

Part of #63008

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

• agent/context_compressor.py: preserve anti-thrash strikes when a verified compaction used the deterministic fallback summary, even if the resulting prompt fits under threshold.
• agent/context_compressor.py: log fallback-marker compactions as degraded attempts through the same ineffective-compression counter.
• tests/agent/test_context_compressor.py: add coverage for fallback strikes, normal successful reset behavior, and the second-strike suppression path.

How to Test

1. Run /Users/leongong/Desktop/LeonProjects/worktrees/hermes-agent/.base/.venv/bin/python -m pytest tests/agent/test_context_compressor.py -q
2. Run /Users/leongong/Desktop/LeonProjects/worktrees/hermes-agent/.base/.venv/bin/python -m py_compile agent/context_compressor.py tests/agent/test_context_compressor.py
3. Run git diff --check

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: macOS 26.4.1

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

Screenshots / Logs

Proof log: /tmp/hermes-63008-proof.txt

[teknium1]

Thanks for the focused regression fix. The R1 premise is present on current main: agent/context_compressor.py:3083 marks a deterministic fallback, while update_from_response() clears the ineffective-attempt count for below-threshold real usage at agent/context_compressor.py:1195 and :1229.

Suggested changes

• Consider one integration-style regression that drives a failed compress() fallback into update_from_response(). The added unit tests cover the intended counter transitions; exercising the existing fallback producer would also preserve the state-machine connection.

No blocking correctness issue identified. This is an automated hermes-sweeper review.

[kshitijk4poor]

Merged via #63965 — your fallback compaction commit was cherry-picked with authorship preserved and extended with a boundary-scoped persisted breaker that survives ordinary fitting responses, usage-less turns, session rotation, and gateway recreation. Thanks for the fix!