| name | read-confluence |
|---|---|
| description | Use when user says "read TDD <link>", "read design doc <link>", or provides a Confluence URL. Fetches the Technical Design Document from Confluence via acli to gather context before implementing. |
Fetch a Technical Design Document (TDD) from Confluence using acli before
starting implementation. The TDD is the source of truth for architecture
decisions, requirements, and design constraints.
Activate when user provides a Confluence page URL, in phrases like:
- "read TDD https://zalora.atlassian.net/wiki/spaces/..."
- "read confluence "
- bare Confluence URL paste
acli must be installed. If a fetch fails because acli is not found, show the
user:
acliis not installed. Install it with:brew install atlassian/tap/atlassian-cliThen authenticate:
acli confluence auth loginRe-run the command once setup is complete.
Do not run which acli before every fetch — only surface this if a command
fails.
Confluence URLs follow this pattern:
https://<domain>/wiki/spaces/<SPACE>/pages/<PAGE_ID>/Page+Title
^^^^^^^^
extract this number
Examples:
https://zalora.atlassian.net/wiki/spaces/FE/pages/123456789/Hero+Banner+TDD→ ID:123456789https://zalora.atlassian.net/wiki/spaces/PLAT/pages/987654321→ ID:987654321
-
Extract the page ID from the provided Confluence URL (the numeric segment after
/pages/). -
Fetch the page:
acli confluence page view --id <PAGE_ID> --body-format view --include-labels
-
Extract and summarize the following:
- Title — document name
- Space — team/area
- Status — current/draft
- Labels — tags/categories
- Body — full content: objectives, design decisions, acceptance criteria, diagrams described in text
-
Present a concise brief:
TDD: <Title> · Space: <SPACE> · Status: <current|draft> Summary: <key objectives and decisions in 3-5 bullet points> Full content: <cleaned-up body> -
Proceed directly to implementation or analysis. Ask one clarifying question only if critical context is missing.
# Fetch page with rendered body + labels
acli confluence page view --id 123456789 --body-format view --include-labels
# Fetch as raw storage format (XML, useful for structured parsing)
acli confluence page view --id 123456789 --body-format storage
# Fetch with child pages listed (useful for multi-page TDDs)
acli confluence page view --id 123456789 --body-format view --include-direct-children
# JSON output
acli confluence page view --id 123456789 --json| Mistake | Fix |
|---|---|
| Asking for design context before reading the doc | Always fetch first, ask after |
Using storage body format |
Use view — it renders human-readable content |
| Ignoring draft status | Check status; drafts may have incomplete specs |
| Missing child pages | Use --include-direct-children for multi-page TDDs |