Hermetiq Skill v0.9.2-beta

Highlights

• Invocation insights & profile-trend support: skill now uses GetInvocationInsights for the curated "what should I change?" list on a single invocation, and GetProfileTrends for cross-build Bazel JSON trace profile analysis. New invocation_insights and profile_trends prompts wired into the MCP prompt list.
• Bottleneck classification glossary: process / analysis / queue / fetch / upload / output-download / cache-check / client-resource bound, each with a recommended drill-down — surfaced in both SKILL.md and the REFERENCE.
• Playbook updates: invocation insights are now step 1 of the slow-build playbook; new "What should I fix in this invocation?" and "Profile trends" rows in the intent-to-tool routing table.
• On-prem MCP URL override: the bundled MCP server URL now reads from HERMETIQ_MCP_URL and falls back to https://mcp.cloud-usc1.hermetiq.io. Cloud users see no change; self-hosted customers can point the plugin at their own deployment via env var with no plugin.json edits.

Hermetiq Skill and MCP Config v0.9.1-beta

Minor updates to the hermetiq skill to reflect recent updates to the Hermetiq MCP server, including group by build ID semantics and improved analytics for build targets and failures. Also, shortened the main guide, clarified build vs invocation workflows, updated tool/prompt guidance, and corrected reference field names to match the current gRPC proto.

Hermetiq Skill and MCP Config v0.9.0-beta

Initial beta release of the Hermetiq AI skill and MCP configuration. The Hermetiq MCP server is in active beta and evolving rapidly.