A late self-cancel can arrive after the RPC service nursery has closed. Starting `Actor.cancel()` there raises before teardown can cancel the shielded parent-channel loop or run lifetime callbacks. Schedule one root-owned task to await `Actor.cancel(None)` before cancelling the root scope. Deduplicate pending requests, and cancel the root scope directly once runtime cancellation has completed. Add a fresh-child regression for open and closed service nurseries under native Trio and asyncio infection. Check repeated requests before completion, after completion and from a lifetime callback; prove all three creator-owned shm segments disappear before parent nursery teardown can mask a failure. Include the session raw record referenced by the structured provenance already in the parent. Prompt-IO: ai/prompt-io/opencode/20260916_ib_disconnect_cleanup.md (this patch was generated in some part by `opencode` using `gpt-6-astra` (`openai`)) |
||
|---|---|---|
| .. | ||
| claude | ||
| opencode | ||
| prompts | ||
| 20261011T053737Z_a966928d_prompt_io.md | ||
| 20261011T053737Z_a966928d_prompt_io.raw.md | ||
| 20261019T140536Z_9a0f220f_final_return_prompt_io.md | ||
| 20261019T140536Z_9a0f220f_final_return_prompt_io.raw.md | ||
| 20261019T145431Z_0553fcf8_ctx_return_fix_prompt_io.md | ||
| 20261019T145431Z_0553fcf8_ctx_return_fix_prompt_io.raw.md | ||
| 20261020T193619Z_0553fcf8_review_prompt_io.md | ||
| 20261020T193619Z_0553fcf8_review_prompt_io.raw.md | ||
| 20261021T145344Z_0553fcf8_task_format_prompt_io.md | ||
| 20261021T145344Z_0553fcf8_task_format_prompt_io.raw.md | ||
| 20261021T151133Z_0553fcf8_uid_move_prompt_io.md | ||
| 20261021T151133Z_0553fcf8_uid_move_prompt_io.raw.md | ||
| 20261028T193020Z_b64c89f0_cached_owner.md | ||
| 20261028T193020Z_b64c89f0_cached_owner.raw.md | ||
| README.md | ||
README.md
AI Prompt I/O Log
This directory tracks prompt inputs and model outputs for AI-assisted development across harnesses.
Policy
Prompt logging follows the NLNet generative AI policy. All substantive AI contributions are logged with:
- Original harness, and provider/model when known
- Timestamps
- The prompts that produced the output
- Unedited model output (
.raw.mdfiles)
Usage
Entries are created by the /prompt-io skill or automatically via /commit-msg integration.
Human contributors remain accountable for all code decisions. AI-generated content is never presented as human-authored work.