Skip to content

server@0.55.0

Choose a tag to compare

@gram-bot gram-bot released this 19 May 21:19
· 4982 commits to main since this release
24f8dde

Minor Changes

  • ecdd727: support remote mcp interceptor payload mutation; implement shadowmcp and mcp:connect interceptors

  • a8cf1e0: Emit audit log entries for collection mutations: mcp_collection:create, :update, :delete, :attach_server, and :detach_server. Update/AttachServer/DetachServer now run in a transaction alongside the audit insert, and a new urn.McpCollection identifier (prefix mcp_collection) is used as the audit subject.

  • 4ea14f3: Enforce RBAC on the collections API. List and ListServers now require org:read; Create, Update, Delete, AttachServer, and DetachServer require org:admin. The dashboard's sidebar, collections list, and detail pages open up to org:read members, while create/edit/delete and server attach/detach controls stay behind org:admin.

  • 5dcb8aa: RiskResult.rule_id and RiskResult.description now follow a consistent shape across every detection source.

    rule_id is lowercase, snake_case, with an optional dot-separated category prefix:

    • secret.<rule> for credentials and secrets (e.g. secret.anthropic_api_key)
    • pii.<rule> for personal, financial, and medical data (e.g. pii.credit_card, pii.medical_license)
    • shadow_mcp for unverified MCP tool calls
    • destructive.tool for MCP tool calls flagged as destructive
    • destructive.<category>.<name> for destructive shell, git, database, and cloud commands (e.g. destructive.shell.rm_rf, destructive.git.push_force)
    • prompt_injection for prompt injection findings

    (source, rule_id) is the stable identifier downstream consumers should match on. The dotted prefix alone is enough to bucket findings by risk category.

    description is a short human-readable sentence describing the finding. It never echoes the matched value and is safe to display verbatim.

    Historical rows written before this release keep their original rule_id and description values; a follow-up migration will rewrite them.

  • 4eadd44: Show assigned roles on pending organization invites and allow org admins to change the role before acceptance. Invite creation and invite role changes now emit audit log entries.

  • 95e1458: The webhooks feature now generates a catalog of event types and schemas for them. This is emitted as an OpenAPI 3.1 document that is synced to svix.

  • 376a74b: Added granular webhook event types for audit log entries — each auditable subject (deployments, projects, MCP servers, API keys, toolsets, risk policies, sessions, and more) now emits its own typed webhook event (e.g. audit_log.deployment_event_v1), enabling subscribers to filter by subject domain rather than receiving all audit activity under a single event type.

Patch Changes

  • bede6e6: Exclude per-request plugin download API key creation from the audit log to prevent flooding with api_key:create events.
  • 4aceb60: skip WorkOS reads when org already linked locally
  • 4eadd44: Invite acceptance now uses Gram invite tokens plus WorkOS User Management Magic Auth codes.
    The server validates the invite token, creates and consumes the Magic Auth code for the invited email, verifies the email match, and completes provisioning.
  • 1562656: Drop Presidio IP_ADDRESS false positives produced from short-form IPv6 strings (b::, dead::, 1::, …) and IPv4 unspecified 0.0.0.0. Analysis of prod risk_results showed these single-hex-group <hex>:: matches dominated IP_ADDRESS noise alongside the existing :: filter; they're now dropped before becoming findings.