Issues

• error-analyzer agent tells the model to call list-errors / error-details, which no PostHog MCP tool or command provides

  PostHog/ai-plugin#247 · Filter by author nabki opened on Oct 3, 2026

• Write gate prompts for approval even in Claude Code bypassPermissions mode

  PostHog/ai-plugin#246 · Filter by author Lorenzo45 opened on Oct 3, 2026

• OpenAI-curated PostHog plugin: OAuth callback consistently returns HTTP 400 in Codex Desktop

  PostHog/ai-plugin#183 · Filter by author wilsonpage opened on Aug 7, 2026
  2

• Skills reference bundled files that are not shipped with them (v1.1.56)

  PostHog/ai-plugin#181 · Filter by author chestap98-lab opened on Aug 6, 2026
  1

• SessionEnd capture omits subagent token usage (Task tool, multi-agent workflows, looped runs) — only the main-session transcript is parsed

  PostHog/ai-plugin#166 · Filter by author madrussa opened on Jul 24, 2026
  1

• Honor a single umbrella opt-in (or documented alias) for LLMA Claude Code capture

  PostHog/ai-plugin#150 · Filter by author gyarra opened on Jul 11, 2026

• Plugin ships ~114 eagerly-listed skills (README says 30+) — please reconsider bundling the full catalog

  PostHog/ai-plugin#147 · Filter by author jokull opened on Jul 10, 2026
  2

• POSTHOG_LLMA_PRIVACY_MODE=true doesn't apply to error

  PostHog/ai-plugin#120 · Filter by author MattGrundy opened on Jun 18, 2026

• Two $ai_generation fidelity bugs in the Claude Code session ingest

  PostHog/ai-plugin#92 · Filter by author mory91 opened on May 26, 2026
  1

• Capture Claude Code sessions when SessionEnd doesn't fire (orchestrated / sandboxed runs)

  PostHog/ai-plugin#77 · Filter by author nkovacic opened on May 13, 2026

• Codex MCP often treats accessible non-active projects as inaccessible instead of switching project

  PostHog/ai-plugin#72 · Filter by author AaronRayRay opened on May 9, 2026

• Support API key authentication for per-project MCP configuration

  PostHog/ai-plugin#4 · Filter by author sabrehagen opened on Mar 24, 2026
  2