Skip to content

A body refused for its size says the largest the endpoint takes, and a service that forwards a command takes none it cannot forward - #113

Closed
riccardone wants to merge 1 commit into
mainfrom
spec-37-413-details
Closed

riccardone wants to merge 1 commit into
mainfrom
spec-37-413-details

Conversation

@riccardone

Copy link
Copy Markdown
Contributor

Entry 40 on #37 (#37), proven by best-peer.

Why

Input Limits has a body too large answered 413, and nothing in the refusal tells the client what would fit, so it cannot correct itself; a body over the limit is refused before it is read, so no field can be named either. A service that forwards a command to another service, in an envelope of its own, can accept one the next hop then refuses after the first said 201. best-peer's 413 carries details.limit_bytes, the most the door that refused takes, and its message names it; a workspace's door takes 4 KiB less than the root, so a command its cloud passes on to a member is never refused there.

What

Input Limits in specs/security.md: details.limit_bytes on a 413, and a forwarding service takes no body it cannot forward whole.

Left out of the entry: whether a schema version may gain bounds in place, as best-peer's 1.0 documents did. That is a versioning question for the owner, not a sentence this PR settles.

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

…a service that forwards a command takes none it cannot forward

Entry 40 of #37. A body too large is 413, with nothing in the refusal a client can act on, and a service that forwards commands may accept one the next hop refuses. best-peer's 413 carries details.limit_bytes, and a workspace's door takes 4 KiB less than the root so a forwarded command always fits.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
riccardone added a commit that referenced this pull request Oct 22, 2026
…a service that forwards a command takes none it cannot forward

Entry 40 on #37 (#37), proven by best-peer. Landed from #113.

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

Copy link
Copy Markdown
Contributor Author

Landed on main as 5af8282 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