atlas-lite-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| ATLASSIAN_URL | No | The URL of your Atlassian Cloud site, for example https://your-site.atlassian.net | |
| ATLASSIAN_EMAIL | No | The email address associated with the Atlassian API token | |
| ATLASSIAN_TOKEN | No | Your Atlassian API token | |
| ATLAS_MCP_CONFIG | No | Absolute path to a different sites.json config file | |
| ATLASSIAN_TOKEN_TYPE | No | The Atlassian token type: classic (the default, for tokens without scopes) or scoped | classic |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| sites_listB | List configured Atlassian sites and what they can reach. |
| page_getB | Read a Confluence page as markdown, storage XHTML, or an outline of headings and tables. |
| page_editB | Exact find and replace on the page storage XHTML. |
| page_update_sectionA | Replace the content under a heading with markdown; keep {{atlas:N}} tokens to preserve macros. |
| page_appendC | Append markdown to the end of a page. |
| page_deleteA | Move a published Confluence page to trash. Does not purge pages or delete drafts. |
| page_createC | Create a page from markdown. |
| page_searchC | Search pages by CQL or plain text. |
| page_commentC | Add a footer comment to a page. |
| table_getC | Read one page table as markdown. |
| table_updateC | Set one table cell, matched by row label and column header. |
| table_add_rowC | Add a table row; cells map column names to values. |
| issue_getB | Read a Jira issue: key fields and description as markdown. |
| issue_searchA | Search issues by JQL or plain text. |
| issue_transitionC | Move an issue to a status. |
| issue_commentC | Comment on an issue in markdown. |
| issue_createC | Create an issue; fields map names or ids to plain values. |
| issue_updateC | Update issue fields, summary or description. |
| issue_deleteA | Permanently delete a Jira issue. Subtasks require deleteSubtasks: true. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 19 tools
Each tool targets a clear resource+action pair, and the three page-mutation variants (page_edit for find/replace, page_update_section for heading-scoped replacement, page_append for end-of-page) are distinguished by explicit descriptions. Minor residual overlap between page_edit and page_update_section could still cause misselection in edge cases, and page_get vs table_get requires the agent to know whether it needs a table specifically.
Names follow a predictable resource_action pattern (page_get, table_update, issue_create, issue_transition) grouped by domain prefix. Small deviations exist: sites_list is plural where others are singular, and multi-word actions like page_update_section and table_add_row break the two-token rhythm slightly, but the convention remains readable.
19 tools serve two distinct product surfaces (Confluence pages/tables and Jira issues), so the count is roughly proportional to scope and each tool maps to a real operation. It sits near the upper end of comfortable, with the page-editing family being the only area that feels finer-grained than strictly necessary.
Both domains have near-full lifecycle coverage: pages support create/read/edit/append/delete/search/comment and issues support get/search/create/update/delete/transition/comment. Gaps are minor but present — no table row deletion or table creation, no page label or attachment operations, and no way to restore a trashed page.