{"repo":"NousResearch/hermes-agent","public_id":"114707","display_label":"Aux summarizer compression cap lost on model switch","summary":"These items concern a bug where the compression threshold for an auxiliary summarizer reverts to the main model's value when the model is switched mid-session, and fixes to ensure the cap is preserved and re-evaluated at switch time.","subsystem":null,"counts":{"members":3,"open_issues":0,"open_issues_distinct":0,"open_prs":0,"closed_issues":1,"merged_prs":1,"attempt_prs":1},"landed_fix":{"status":"merged_to_main","basis":"structural","pr":114823,"merged_at":"2026-09-18T16:49:15+00:00"},"resolution":{"state":"resolved","by":"pr-114823"},"freshness":{"state":"verified","last_verified_at":"2026-09-18T16:49:15+00:00","is_stale":false},"grouping":{"backbone":"STRUCTURAL_SPANNING","link_density":0.6666666666666666},"fix_edges":[{"issue_number":114707,"pr_number":114823,"tier":"verified","basis":"closing_ref","confidence":0.95,"recoverable_from_timeline":true,"pr_merged_at":"2026-09-18T16:49:15+00:00"}],"members":[{"kind":"issue","number":114707,"title":"Compression threshold reverts to main-model value after mid-session model switch, leaving auxiliary summarizer undersized","url":null,"state":"closed"},{"kind":"pr","number":114710,"title":"fix(agent): preserve auxiliary compression cap after model switch","url":null,"state":"closed","close_reason":{"refs":[{"kind":"commit","value":"a37293a2ceda","number":0},{"kind":"pr","value":"#114823","number":114823},{"kind":"pr","value":"#114707","number":114707},{"kind":"pr","value":"#114710","number":114710}],"category":"merged_via","implication":"superseded"},"merged":false},{"kind":"pr","number":114823,"title":"fix(compression): aux summariser clamp survives model switches and is re-checked at switch time (#114707, salvage #114710)","url":null,"state":"merged","close_reason":{"category":"merged"},"merged":true,"is_fix_pr":true,"fix_tier":"verified","files_changed":["agent/agent_runtime_helpers.py","agent/chat_completion_helpers.py","agent/context_compressor.py","agent/conversation_compression.py","agent/turn_preflight.py","tests/agent/test_compression_feasibility.py","tests/agent/test_primary_runtime_restore.py","tests/agent/test_switch_model_context.py","website/docs/developer-guide/context-compression-and-caching.md"]}],"trust":{"default_tier":"verified","includes_suggested":false}}