Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
Python agentaianthropicclaude-codecompressioncontext-engineeringcontext-windowcursorfastapilangchainllmmcpopenaiprompt-engineeringproxypythonragtoken-optimizationtokenstypescript
fix(smart_crusher): drop pre-fix CCR override hack — use dataclass defaults
chopratejasca2cb450c59b
diff --git a/headroom/transforms/smart_crusher.py b/headroom/transforms/smart_crusher.pyindex 00eea05d3..3902f5ff3 100644--- a/headroom/transforms/smart_crusher.py+++ b/headroom/transforms/smart_crusher.py@@ -188,12 +188,21 @@ class SmartCrusher(Transform):# reference; storing it under `enabled=False` would also be a# surprise side effect the user explicitly disabled).#+ # Default falls through to `CCRConfig()` so direct callers+ # (the proxy and tests that don't pass an explicit config) get+ # the documented dataclass defaults (`enabled=True,+ # inject_retrieval_marker=True`). The previous override here+ # set `inject_retrieval_marker=False` as a no-op-intent hack+ # back when the Rust port silently ignored the flag; now that+ # the flag is honored, that override would actively suppress+ # markers + store writes for every caller.+ ## Scope: gates ONLY the row-drop sentinel path. Stage-3c.2# opaque-string CCR substitutions still emit always — they have# no Python equivalent and no production caller has asked for# them to be suppressed.if ccr_config is None:- self._ccr_config = CCRConfig(enabled=True, inject_retrieval_marker=False)+ self._ccr_config = CCRConfig()else:self._ccr_config = ccr_config