Compaction as Epistemic Failure: How Agentic LLM Tools Fabricate Confirmed Results from Killed Processes
On this page
arXiv:2607.13071v2 Announce Type: replace-cross Abstract: Agentic LLM coding tools compress long session histories into compaction summaries that subsequent sessions inherit as ground truth. This paper documents a failure mode in Claude Code where partial standard output from timed-out commands (exit code 143) is recorded in compaction summaries as confirmed results, propagating false positives across sessions and model versions without re-verification. The underlying mechanism is a conflation of observation and persistence, where information that appeared in the terminal is treated as equivalent to information written to durable storage. This finding extends the analysis of LLM self-evaluation failures reported in prior work on non-determinism in LLM-as-judge grading by showing that agentic tools exhibit analogous reliability deficits when reporting on their own operational outcomes. The failure has direct implications for any workflow that relies on agentic session continuity for data processing, scientific computation, or multi-step automation.
Why It Matters
Agentic LLM coding tools compress long session histories into compaction summaries that subsequent sessions rely on. This paper shows how compaction can become an epistemic failure—fabricating confirmed results from processes that were actually killed or never completed. It matters because agents that confidently report unverified outcomes corrupt downstream work and trust. See Agents, Tools, and Agent Frameworks, Agents, Autonomy, and Tool Use, AI Safety, and Knowledge, Reasoning, and Semantics.
Sources
Back to Recent News · Sub-article index.