vultr/openclaw
1
0
Fork 0
mirror of https://github.com/openclaw/openclaw.git synced 2026-05-06 16:50:43 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(openai-codex): avoid stale Responses replay state

Browse Source
This commit is contained in:
mkdev11
2026-05-03 19:05:23 +02:00
committed by Peter Steinberger
parent 8285786c22
commit 27e467ad23
6 changed files with 273 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
docs/reference/transcript-hygiene.md
View File

@@ -117,6 +117,7 @@ inter-session user turns that only have provenance metadata.
- Image sanitization only.
- Drop orphaned reasoning signatures (standalone reasoning items without a following content block) for OpenAI Responses/Codex transcripts, and drop replayable OpenAI reasoning after a model route switch.
- Preserve replayable OpenAI Responses reasoning item payloads, including encrypted empty-summary items, so manual/WebSocket replay keeps required `rs_*` state paired with assistant output items.
- Native ChatGPT Codex Responses is the exception: OpenClaw does not replay prior Responses reasoning/message/function item IDs or session `prompt_cache_key` to avoid stale backend replay across turns.
- No tool call id sanitization.
- Tool result pairing repair may move real matched outputs and synthesize Codex-style `aborted` outputs for missing tool calls.
- No turn validation or reordering.