confluence-publisherlisted

# Confluence Publisher Publish a single page to Confluence — create or update — from Markdown, storage XHTML, or plain text. Companion to `confluence-crawler`: same credentials namespace, same flavor support, opposite direction. ## Instructions You are a Confluence publishing agent. Authentication, REST mechanics, optimistic-locking retries, and the Markdown→storage conversion live in `scripts/`. Do not re-implement any of that; invoke the script with the right flags and report the result. ### Flavor support Same as the crawler: - **Atlassian Cloud** (`*.atlassian.net`) — Basic auth with email + API token. Base URL must include `/wiki`. - **Confluence Server / Data Center** — Bearer auth with a Personal Access Token. Flavor is auto-detected from the base URL; override via `CONFLUENCE_FLAVOR=cloud|server` if needed. ### Configuration location Credentials are resolved by the build-projected `credentials_shim.load_credentials` through Tier 1 (env) → Tier 2 (OS keyring) → Tier 3 dotfile. The dotfile lives at `~/.agentbundle/credentials.env`. The declared schema is at `references/creds-schema.toml` and shares the `confluence` namespace with `confluence-crawler` — if either skill has been configured, this one works. | Key | Required | Notes | |---|---|---| | `CONFLUENCE_BASE_URL` | yes | Cloud: `https://<site>.atlassian.net/wiki`. Server: `https://confluence.corp.example.com`. | | `CONFLUENCE_API_TOKEN` | yes | Cloud API token or Server PAT. | | `CONFLUENCE_EMAIL` | C