Skip to content

A command field marked writeOnly is taken at admission and never recorded, replayed or emitted, and its command is not run again without it - #116

Closed
riccardone wants to merge 1 commit into
mainfrom
spec-37-write-only-fields
Closed

riccardone wants to merge 1 commit into
mainfrom
spec-37-write-only-fields

Conversation

@riccardone

Copy link
Copy Markdown
Contributor

Entry 50 on #37 (#37), proven by best-peer.

Why

The spec records every accepted command whole, so it can be replayed and run again after a restart (#102). Nothing says what a service does with a field that must never be kept: a credential, or a URL that grants access by itself. best-peer: a command's data schema marks such a field writeOnly; the admission check takes it and keeps it where it belongs, such as a file of the workspace readable by the service alone; the command is then recorded, and handed to its handler, without it, so neither the log, a replay nor an event holds it again. After a restart, a command owed whose recorded copy lacks what was sent is failed with the reason, "send it again", never run without its secret.

What

Ingestion Semantics gains a paragraph after its bullets, placed so it does not touch the bullets #88, #92 and #102 change.

Nothing changes in the conformance list: the sentence is a clarification.

No version bump and no release notes here: the PRs from #37 do not collide that way, and the release that gathers them does both.

🤖 Generated with Claude Code

…rded, replayed or emitted, and its command is not run again without it

Entry 50 of #37. The spec records every accepted command whole, so it can be replayed and run again after a restart, and says nothing of a field that must never be kept: a credential, or a URL that grants access by itself. best-peer's schemas mark such a field writeOnly, the admission keeps it where it belongs, the command is recorded without it, and after a restart such a command is failed rather than run without it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
riccardone added a commit that referenced this pull request Oct 22, 2026
…rded, replayed or emitted, and its command is not run again without it

Entry 50 on #37 (#37), proven by best-peer. Landed from #116.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@riccardone

Copy link
Copy Markdown
Contributor Author

Landed on main as fa65eb7 under the repository's new process (SDLC.md: every change on main after the local gate and the review). The review of the whole batch is in the Review: commits after it, up to 5603cb1. Closing this pull request.

@riccardone riccardone closed this Oct 22, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant