docs(spec): Claude Desktop coverage strategy#175
Draft
7schmiede wants to merge 2 commits into
Draft
Conversation
Development
Honest coverage matrix for Claude Desktop on Pro/Max given Anthropic's 2026 policy posture (Feb 2026 Authentication and Credential Use, April 2026 third-party harness enforcement). Three-tab data flow analysis (Chat/Cowork/Code), four-component defence-in-depth (MCP Connector, Audit Mirror, Endpoint DLP, new Pre-flight tool), and a decision record on what we deliberately do not build. Pre-flight is the only new component: a small Tauri-based client that helps users pseudonymise text BEFORE pasting into Claude Desktop and resolve pseudonyms AFTER copying responses back. Does not interact with Claude Desktop itself in any way; ToS-clean by design. Supersedes the closed Pseudonym Bridge spec (PR #174), which targeted claude.ai chat webview — wrong abstraction for the customer's actual need (Code tab coding-agent governance). Status: draft, not yet approved. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
7 tasks
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Honest coverage spec for the customer constraint that came out of the PR #174 discussion: Pro/Max stays, Claude Desktop stays (especially the Code tab), API mode is not a viable alternative. The Pseudonym Bridge approach in #174 targeted the wrong abstraction (claude.ai chat) and Anthropic's 2026 policy posture (Feb Authentication and Credential Use + April third-party harness enforcement) closes any wrap-or-proxy approach for Pro/Max.
This spec defines what Powerbrain can defensibly deliver inside those constraints — a four-component defence-in-depth — and what it deliberately does not build.
Spec doc: docs/specs/2026-05-22-claude-desktop-coverage-strategy.md
What changed since #174
The four components
Pre-flight is the only buildable new component. ToS-clean by design: it's a clipboard / text utility that calls Powerbrain APIs, like a password manager that runs alongside your browser.
What we deliberately do not build (decision record)
Captured in the spec so the question doesn't return:
What customers can honestly claim
Spec includes a defensibility matrix — what claims are substantiated by which component, and an explicit list of inflated claims to avoid. This is what compliance teams need for DPO/DPA review.
What changed in this PR
No code changes. Doc-only.
Open questions in the spec
Test plan
Doc-only.
Checklist