docs(reference): add the protocol fidelity matrix - #490
Merged
Conversation
Output of the fidelity sweep prompted by the competitive analysis: a per-path × per-dimension table of what grob preserves, transforms, cannot carry (API-limited) or intentionally drops when translating between the Anthropic canonical format and each provider wire format. Makes the losses explicit rather than silent — the analysis's core critique of gateways is that a translation loss looks like HTTP 200. Records the fixes landed this cycle (#476 message images, #488 tool-result images on Responses, #465 reasoning continuity, #489 structured upstream errors) and flags the unverified rows (Chat/Gemini tool-result images) honestly, since grob has no standing test credential for those backends. Feeds the per-release conformance-suite work (#484). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Destynova2
enabled auto-merge (squash)
July 28, 2026 21:09
Merged
Destynova2
added a commit
that referenced
this pull request
Aug 4, 2026
## 🤖 New release * `grob`: 0.36.85 -> 0.36.86 <details><summary><i><b>Changelog</b></i></summary><p> <blockquote> ## [0.36.86](v0.36.85...v0.36.86) - 2026-08-04 ### Fixed - *(policies)* enforce decision token expiry and audience ([#500](#500)) - harden critical architecture boundaries ([#493](#493)) ### Other - *(design)* validate the OCR-to-DLP hypothesis against the real engine ([#499](#499)) - *(design)* measure L2 watermark robustness, correct layer table ([#498](#498)) - *(design)* measure TrustMark cost and pHash robustness contract ([#497](#497)) - *(design)* correct c2pa cost with real measurements, reorder media plan ([#496](#496)) - *(design)* image/video DLP, content provenance, and agent control plane ([#495](#495)) - *(reference)* add the protocol fidelity matrix ([#490](#490)) </blockquote> </p></details> --- This PR was generated with [release-plz](https://github.com/release-plz/release-plz/).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The concrete artifact from the fidelity sweep: a per-path × per-dimension table
of what grob preserves, translates, cannot carry (API-limited), or intentionally
drops between the Anthropic canonical format and each provider wire format.
Makes each loss explicit rather than silent — the analysis's core critique is
that a translation loss looks like
HTTP 200. Records this cycle's fixes andflags the unverified rows honestly.
Findings
upstream_error(#489)The Responses tool-result-image fix (#488) was verified live; the Chat/Gemini
equivalents are API-shaped losses (string-only tool content) marked unverified
because grob has no standing test credential for those backends.
Feeds the per-release conformance-suite work (#484).
🤖 Generated with Claude Code