History leakage: when an agent's edit history pollutes its context — and a /dehistorize skill to fix it
dmkii · reddit · 2026-10-10
The author identifies a common problem: history leakage. An agent's edit history seeps into everything — models desperately trying to prove themselves or oversharing what was already deleted, leaving context littered with artifacts like ScoringV2 and legacyphase. It mirrors how humans anchor to a number or let a delta influence decisions instead of taking a fresh look.
The fix is a skill called /dehistorize (open-sourced in a GitHub agent-skills repo): a single word wipes out those historical residues, so the UI can support the next decision instead of reporting the agent's past work.
More from coding & agent
- Anthropic MAX plans now include matching monthly API credits; one Opus 5.5 prompt ported his Agent SDK side project to Managed Agents — trq212 · 2026-10-10
- Nace AI's sub-10B decision model Drex v1.5 lands on OpenRouter — nischay_twt · 2026-10-10
- Command Center 1.3.0 ships with 40+ improvements for understanding your agent systems — jimmykoppel · 2026-10-10
- Agents kept 'worsen-bettering' her setup — one essay folder hit 862 files, 1M words — every · 2026-10-10
- Theo, zeeg and other founders on agents: they ship code faster, but have no talent or curiosity — zeeg · 2026-10-10
- Celesto Computer MCP now listed on Smithery marketplace — aniketmaurya · 2026-10-10