jira-taskslisted

# Purpose Read Jira to answer "what am I working on", "what's the status of `<PROJECT-KEY>-115`", or to reconcile workspace tasks against tickets. **Read-only by default** — every write (create / transition / comment / worklog) is a separate, explicit, user-requested action (see [Write-op policy](#write-op-policy)). **Site and project key:** this skill uses `<jira-base-url>` (e.g. `yourteam.atlassian.net`) and `<PROJECT-KEY>` as placeholders. Check the consumer project's `CLAUDE.md` (or `.claude/project.json`) for the actual Jira base URL, project key, and any pinned cloudId. If none is documented, ask the user once and suggest recording it in `CLAUDE.md`. # Tool flow The Atlassian MCP tools are **deferred** — their schemas are not loaded at start. Load them before the first call, or the call fails with `InputValidationError` (the tool-name prefix depends on how the Atlassian MCP server is registered on the host): ``` ToolSearch query: "select:mcp__plugin_atlassian_atlassian__getAccessibleAtlassianResources,mcp__plugin_atlassian_atlassian__searchJiraIssuesUsingJql,mcp__plugin_atlassian_atlassian__getJiraIssue" ``` Every Jira call needs a **cloudId**. Resolve it by calling `getAccessibleAtlassianResources` (no args) and taking the `id` of the `<jira-base-url>` resource. If the project's `CLAUDE.md` pins a cloudId, use it directly — but **if any call 404s** (site migrated, token re-scoped), re-resolve rather than retrying a stale id. - **Search / list** → `searchJiraIss