confluence-publish

Publish, create, and update Confluence pages from HTML content.

Security Scan

The skill's code, runtime instructions, and requested environment access are consistent with its stated purpose of publishing/upserting Confluence pages and do not request unrelated credentials or network endpoints.

Purpose & Capability

Name and description match the actual behavior: the skill upserts Confluence pages via the Confluence REST API. Requested credentials (EMAIL, DOMAIN, API_TOKEN), workspace filesystem read, and network access restricted to *.atlassian.net are appropriate and expected.

Instruction Scope

SKILL.md and src/main.py only instruct the agent to read page input (from the provided input string or a workspace file), optional .env-style files under the workspace, and to call Confluence API endpoints on Atlassian Cloud. There are no instructions to read unrelated system files, extra env vars, or to contact external endpoints outside *.atlassian.net.

Install Mechanism

This is an instruction-and-python-source skill with no install spec; requirements.txt only lists requests and pytest. No remote downloads, extract steps, or unusual install behaviors are present.

Credentials

The skill requires exactly EMAIL, DOMAIN, and API_TOKEN (API_TOKEN marked primary), which are the minimal credentials needed to authenticate with Confluence. It does not request unrelated secrets or other service credentials.

Persistence & Privilege

The skill is not always-on, does not modify other skills, and requests only read access to files under the workspace and network access to Atlassian Cloud. Autonomous invocation is allowed (platform default) but is not combined with broad or unrelated privileges.

Assessment

This skill appears coherent and implements safe checks (validates that DOMAIN targets *.atlassian.net, enforces workspace-only file reads). Before installing: provide only a Confluence API token with the minimum required privileges, ensure any env_file or page_path you pass is inside your workspace (the skill rejects outside files), and review the included src/main.py if you want to confirm no additional network endpoints are added. Consider testing in an isolated environment or with a test Confluence space/account first.

Publish, create, and update Confluence pages from HTML content.

What this skill does

This skill upserts Confluence pages through the Confluence REST API:

• If a page with the same title exists in the target space, it updates the page.
• If no page exists, it creates a new page.

It supports credentials from config, env file, or process environment variables.

Actions

publish_page

Create or update a Confluence page.

Expected input:

• input: HTML content, optionally with metadata JSON in the first HTML comment.
• config: runtime options and credentials.

Metadata comment format:

<!--
{"space_key":"SPACE","page_title":"My Page","parent_page_id":"12345"}
-->
<h1>Body content</h1>

Alternative config keys:

• space_key
• page_title
• parent_page_id (optional)
• body_html
• page_path (path to a file in the current workspace containing metadata comment + body)

Credential options:

• config.credentials.EMAIL, config.credentials.DOMAIN, config.credentials.API_TOKEN
• config.env_file pointing to a .env style file in the current workspace
• Environment vars: EMAIL, DOMAIN, API_TOKEN

Security constraints:

• config.env_file and config.page_path must resolve to files under the current workspace directory.
• DOMAIN and optional base_url must target Atlassian Cloud (https://<tenant>.atlassian.net/wiki).

Success output includes:

• status: "success"
• operation: "created" | "updated"
• page_id
• title
• url
• space_key

test_connection

Checks Confluence authentication and returns user identity info.

Example call payload

{
  "action": "publish_page",
  "input": "<!-- {\"space_key\":\"SPACE\",\"page_title\":\"Demo\"} --><h1>Hello</h1>",
  "config": {
    "credentials": {
      "EMAIL": "[e-mail]",
      "DOMAIN": "exampletenant",
      "API_TOKEN": "your-token"
    }
  }
}