Skip to content

History / Security Proof Client Credential Isolation

Revisions

  • docs: correct stale/security-relevant claims found in a full wiki audit - Security-Proof-Client-Credential-Isolation.md: the 'Client isolation matrix' still described DCR redirect-URI scope inheritance as current, safe behavior -- that was the exact #497 P0 vulnerability, fixed in #505. Corrected the table and added a note. - Home.md, Scope-and-Tool-Matrix.md, OAuth-Clients-and-Scopes.md: all three said anonymous/read access requires 'no auth at all', which is only true with OAuth disabled. Production runs with OAuth enabled, where every /mcp request without a bearer is challenged regardless of tool tier (#498, confirmed live). Added correction notes distinguishing the scope-ACL layer (genuinely ungated) from bearer-token requirement (not bearerless in production). - AgentReady-Pitfalls.md, Live-Deployment-Runbook.md: documented a second, distinct cause of the 'chtimes ... operation not permitted' build failure (#521, live-reproduced 2026-07-18) -- the MCP server's own build_site tool runs as a different system user (mcp-hugo-server-go) than the manual rebuild path these pages already documented (hugo-mcp), and can be blocked by individual files that drifted to the wrong owner even when ReadWritePaths is already correct.

    @jmrGrav jmrGrav committed Jul 18, 2026
  • docs: update wiki to reflect the #450 2-scope model (read/write) Several pages still described the pre-#450 four-tier scope model (content.read/content.write/site.admin/system.admin with mcp as the only alias) as if it were still active. Since #450 the server enforces exactly two scopes, read and write, with every older scope string accepted only as a compatibility alias. Home, Documentation-Source-of-Truth, OAuth-Clients-and-Scopes, Scope-and-Tool-Matrix, Pitfalls-Index, Client-Compatibility, and Smoke-Tests-and-Live-Verification updated to describe the current model as current, with the old model kept only as historical context where relevant. Pitfall-Anonymous-vs-Authenticated-Server rewritten more substantially: it described a four-MCP-server-instance architecture (anonServer/readServer/writeServer/siteAdminServer) that no longer exists in the code -- the server now runs two instances (publicServer/writeServer), and read-only tools that used to require content.read (validate_site, build_agent_context, etc.) are fully ungated since #450, not just relabeled. Security-Proof-Client-Credential-Isolation gets a terminology note rather than a rewrite -- the security property it demonstrates (redirect URI isolation blocks cross-client credential reuse) is scope-model-independent, so the historical scope names are left in place with a pointer to the current model.

    @jmrGrav jmrGrav committed Jul 18, 2026
  • docs: add security proof for client credential isolation Live test (2026-07-05): using chatgpt-write credentials from Claude.ai is refused because the redirect_uri sent by Claude.ai does not match the URIs registered for the chatgpt-write client. Documents the cross-client token theft protection as a concrete proof with RFC 6749 reference. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

    @jmrGrav jmrGrav committed Jul 5, 2026