[dependabot]

Bumps github.com/posthog/posthog-go from 0.0.0-20230801140217-d607812dee69 to 1.42.0.

Release notes

Sourced from github.com/posthog/posthog-go's releases.

1.42.0

Unreleased

1.41.1

Unreleased

1.41.0

Unreleased

1.40.1

Unreleased

1.40.0

Unreleased

1.39.0

Unreleased

1.38.0

Unreleased

1.37.0

Unreleased

1.36.2

Unreleased

1.36.1

Unreleased

1.36.0

Unreleased

1.35.3

Unreleased

1.35.2

Unreleased

1.35.1

Unreleased

1.35.0

Unreleased

1.34.0

Unreleased

1.33.0

Unreleased

... (truncated)

Changelog

Sourced from github.com/posthog/posthog-go's changelog.

1.42.0

Minor Changes

• 761d995: Add posthogmcp.WithCaptureErrorMessages(false), which keeps a failed MCP tool call's error text out of $mcp_error_message and the $exception event. Tool errors often repeat their arguments, and only credentials are redacted from them. The failure is still recorded with its error type, and both carry Tool <name> returned an error instead. It applies to the manual API and to the posthogmcpsdk and posthogmcpgo adapters. An IsError result's content is still part of $mcp_response, so turn response capture off too to keep failure text out entirely.

Patch Changes

• a528b3f: Aligned posthogmcpsdk with the sdk-specs MCP spec. The get_more_tools virtual tool now advertises readOnlyHint, idempotentHint, destructiveHint and openWorldHint as false, since it records data, and no longer sets a title. The injected context argument has a shorter description that asks for a few words on why the tool is called, described by role. The injected llm_model argument is no longer listed as required. This changes your published tool definitions; if your server is listed in the ChatGPT or Claude directories, those platforms may re-review the updated tools.

1.41.1

Patch Changes

• 17365fe: Suppress local-evaluation definition warnings for client-only flags, including definitions loaded from a shared cache, without changing their evaluation behavior.

1.41.0

Minor Changes

• d4ea159: Add posthogmcpgo, MCP analytics for servers built on github.com/mark3labs/mcp-go. posthogmcpgo.Instrument returns one server.ServerOption that captures $mcp_tool_call events with the same properties as posthogmcpsdk, including the agent's intent from an added context argument. Tools whose input schema disallows other properties get no added arguments, so servers using WithInputSchemaValidation keep working.

1.40.1

Patch Changes

• 8fef83f: Reworded the descriptions of the injected conversation_id and llm_model arguments and the get_more_tools virtual tool so they describe the field instead of instructing the model. This changes your published tool definitions; if your server is listed in the ChatGPT or Claude directories, those platforms may re-review the updated tools.

1.40.0

Minor Changes

• 03e1647: The new Client.Emails() sends transactional email from your Go code and reads each email's delivery status.

  Breaking for custom clients. A type that implements Client itself now needs an Emails method.

  client, err := posthog.NewWithConfig("phc_...", posthog.Config{SecretKey: "phs_..."})
  if err != nil {
  	log.Fatal(err)
  }
  defer client.Close()
  result, err := client.Emails().Send(ctx, posthog.CodeEmail{
  To:          "user@example.com",
  DistinctID:  "user_123",
  TemplateKey: "password-reset",
  Subject:     "Reset your password",
  HTML:        renderedHTML,
  })

... (truncated)

Commits

• See full diff in compare view

[snyk-io]

⛔ Snyk checks have failed. 3 issues have been found so far.

Status	Scan Engine	Critical	High	Medium	Low	Total (3)
⛔	Open Source Security	0	3	0	0	3 issues
✅	Code Security	0	0	0	0	0 issues

💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse.