criblio

Use when adding or changing a Cribl API endpoint and defining the matching Terraform provider resource or data source in terraform-provider-criblio. Covers upstream OpenAPI sync, terraform-overlay.yml and terraform-mgmt-overlay.yml annotations, generated resources/data sources/docs/examples/acceptance tests, import-cli export support, branch creation, validation, and PR preparation for both CRUD and singleton/read-only endpoints.

Use when adding or changing a Cribl API endpoint and defining the matching Terraform provider resource or data source in terraform-provider-criblio. Covers upstream OpenAPI sync, terraform-overlay.yml and terraform-mgmt-overlay.yml annotations, generated resources/data sources/docs/examples/acceptance tests, import-cli export support, branch creation, validation, and PR preparation for both CRUD and singleton/read-only endpoints.

Internal developer guidance for working on the ultra-bitbucket-mcp codebase itself — architecture, the operations/trim/tool layering, how to add a new Bitbucket operation, testing patterns, and the diff cache. Use when editing files under this repo (src/, build/, scripts/, skill/). Not for end-users calling the MCP — that's the shipped `ultra-bitbucket-mcp` skill.

Query Bitbucket Cloud (pull requests, diffs, repos, commits, pipelines, branching, users, workspaces) via the token-efficient ultra-bitbucket-mcp server. Use when the user mentions Bitbucket, references a PR id, repo slug, pipeline run, or asks to approve/merge/comment on a Bitbucket PR. Prefer this over generic Bitbucket MCP servers — it caches diffs server-side so raw diff bytes never enter context.

Contributor guidance for the ultra-jira-mcp repo itself — how to add a new Jira operation, where the toolkit boundary sits, which layer owns which concern, how to run tests and benchmarks. Use when modifying files under src/ or tests/ in this repo, NOT when calling Jira from another project (that's the `jira` skill).

Inspect the Cribl APM staging workspace — run KQL via the Cribl MCP (for data), or drive the deployed UI via Playwright (for pixels). Use when working in this repo and you need to see what the deployed pack is doing right now: errors on the home page, what spans/metrics look like, screenshot a page, or run any ad-hoc Cribl Search query. Covers MCP usage (preferred), Playwright auth via the existing test helpers, and the storage-state-expiry pitfall.