Confluence Skill

Interact with Confluence Cloud using the confluence.sh script at ~/.claude/skills/confluence/scripts/confluence.sh. All calls go through the Confluence REST API (v2 primarily, v1 for CQL search) using curl with Basic Auth.

Prerequisites

Same Atlassian credentials as the Jira skill. The user must have these environment variables set (typically in ~/.zshrc):

JIRA_DOMAIN or JIRA_URL   - e.g. mycompany.atlassian.net
JIRA_EMAIL                 - e.g. user@example.com
JIRA_API_TOKEN             - from https://id.atlassian.com/manage-profile/security/api-tokens

If any are missing, tell the user which variables need to be set and link to the token page.

How to Use

Run commands via Bash:

bash ~/.claude/skills/confluence/scripts/confluence.sh <command> [args...]

Available Commands

Command	Args	Description
test		Verify authentication works
get-spaces	[limit]	List available spaces
get-pages	<SPACE_ID> [limit] [sort]	List pages in a space
get-page	<PAGE_ID> [format]	Get page content
search	<CQL> [limit]	Search with CQL
create-page	<SPACE_ID> <TITLE> <BODY> [PARENT_ID]	Create a page
update-page	<PAGE_ID> <TITLE> <BODY>	Update a page
get-child-pages	<PAGE_ID> [limit]	List child pages
get-comments	<PAGE_ID> [limit]	Get footer comments
add-comment	<PAGE_ID> <BODY>	Add a footer comment
get-labels	<PAGE_ID>	Get page labels
add-label	<PAGE_ID> <LABEL>	Add a label

Workflow Patterns

Finding a page: First get the space ID, then list pages or search.

bash ~/.claude/skills/confluence/scripts/confluence.sh get-spaces
bash ~/.claude/skills/confluence/scripts/confluence.sh get-pages 123456

Reading a page: Get page by ID with storage format (default) or ADF.

bash ~/.claude/skills/confluence/scripts/confluence.sh get-page 789012

Searching content: Use CQL (Confluence Query Language) for flexible search.

bash ~/.claude/skills/confluence/scripts/confluence.sh search "type = page AND space = 'PROJ' AND text ~ 'deployment'"

Creating a page: Provide space ID, title, and HTML body in Confluence storage format.

bash ~/.claude/skills/confluence/scripts/confluence.sh create-page 123456 "Page Title" "<p>Content here</p>"

Updating a page: Version number is auto-incremented by the script.

bash ~/.claude/skills/confluence/scripts/confluence.sh update-page 789012 "Updated Title" "<p>New content</p>"

Output Handling

• Pipe output through jq for formatting when presenting to the user
• For get-spaces, present as a table with ID, key, and name
• For get-page, extract title, body content, and version info
• Storage format body contains XHTML - convert to readable text when presenting
• For search, extract and present title, space, and page ID from results

Content Format

Confluence uses XHTML-like "storage format" for page bodies. For simple content, wrap text in <p> tags. For rich content (code blocks, panels, tables), consult ~/.claude/skills/confluence/references/confluence-api.md.

Error Handling

• 401: Authentication failed - check JIRA_EMAIL and JIRA_API_TOKEN
• 403: Permission denied - user may not have access to that space/page
• 404: Page or space not found - verify the ID
• 409: Version conflict - page was modified by someone else (retry)
• 429: Rate limited - wait a few seconds and retry

Reference

For detailed API endpoint documentation, CQL syntax, storage format examples, and macros, see ~/.claude/skills/confluence/references/confluence-api.md.