# Twg Confluence

> >

- **Type:** Skill
- **Install:** `agentstack add skill-atlassian-twg-cli-twg-confluence`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [atlassian](https://agentstack.voostack.com/s/atlassian)
- **Installs:** 0
- **Category:** [Productivity](https://agentstack.voostack.com/c/productivity)
- **Latest version:** 0.1.0
- **License:** Apache-2.0
- **Upstream author:** [atlassian](https://github.com/atlassian)
- **Source:** https://github.com/atlassian/twg-cli/tree/main/skills/twg-confluence
- **Website:** https://teamwork-graph.atlassian.com/cli/install

## Install

```sh
agentstack add skill-atlassian-twg-cli-twg-confluence
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# twg-confluence

Use with the root `twg` skill whenever Confluence is the source of truth or a
Confluence mutation is required. This skill owns content-type, hierarchy,
format, and concurrency semantics; exact command grammar comes from live help.

## CLI launcher fallback

Run `twg `. On shell `command not found`, use `$HOME/.local/bin/twg`
(macOS/Linux) / `$env:LOCALAPPDATA\Programs\twg\bin\twg.exe` (PowerShell), then
tell user to add that directory to PATH. Do not treat auth or command errors as
PATH failures.

## Use When

- A Confluence URL, content ID, space key, page title, or CQL expression is the
  primary anchor.
- The user asks to create, edit, move, copy, comment on, export, archive, or
  organize Confluence content.
- A workflow needs authoritative page bodies, hierarchy, versions,
  permissions, or space metadata.

Do not load this skill merely because a Confluence page is one supporting link
inside a broader workflow.

## First Route

| Intent                           | Route                                                                                  |
| -------------------------------- | -------------------------------------------------------------------------------------- |
| Known content ID or URL          | `confluence content get`                                                               |
| Exact Confluence filtering       | Confluence search with CQL                                                             |
| Date-filtered page/blogpost list | `confluence search query --cql` with `lastmodified`; `content list` has no date filter |
| Fuzzy page/topic discovery       | Cross-product search, then native get                                                  |
| Create or update content         | Unified `confluence content` surface                                                   |
| Space metadata/lifecycle         | `confluence space`                                                                     |
| Hierarchy                        | `confluence tree`                                                                      |
| Export                           | Word returns a download directly; PDF requires export-status polling                   |

Use `twg help describe ""` before an unfamiliar or consequential
mutation.

## Confluence Semantics

- Use the unified content surface for pages, blog posts, live docs,
  whiteboards, databases, folders, comments, versions, history, permissions,
  tasks, labels, and attachments when advertised.
- For Confluence Share dialog access changes, treat General access as
  `restriction-state` and Specific access rows as direct `permissions`:
  `Open / Can edit` -> `OPEN`, `Open / Can view` -> `EDIT_RESTRICTED`,
  `Restricted` -> `VIEW_RESTRICTED`; Specific `Can edit` -> `update`, and
  Specific `Can view` -> `read`.
- Use `confluence space` for spaces and `confluence tree` for hierarchy.
- A stable URL may resolve directly, but exact help remains authoritative about
  accepted ID, URL, site, and content-type forms.
- Search snippets are discovery candidates. Read the selected content before
  summarizing or editing it.
- Page titles are supplied separately from bodies. Do not repeat the title as
  the first body heading.

## Safe Authoring And Editing

- Treat user authorization and CLI confirmation as separate checks for archive,
  trash, and purge. The CLI requirement to pass `--yes` is an execution
  preflight, not evidence that the user authorized the mutation. Use `--yes`
  only after the user explicitly named the mutation and the exact target or a
  previously displayed bounded target set.
- Explicit instructions such as "archive page XYZ" or "archive those 15 pages"
  after a concrete list authorize execution without another confirmation.
  Vague outcome language such as "clean up", "organize", "remove clutter", or
  "take care of" does not authorize archive, trash, or purge.
- If the mutation or targets were inferred, read and resolve the proposed
  targets, present the exact action and bounded target list, and wait for the
  user to affirm that mutation. Do not generalize approval from a similar prior
  action. `trash` does not authorize an additional permanent purge.
- Before creating or editing content in a space, read that space's instructions
  with `confluence space instructions get --key ` (or with
  `confluence space instructions get --id `) and apply them while
  authoring; see `references/spaces.md`. Absent or empty ⇒ author with defaults.
- For an unspecified new internal "page," "doc," "write-up," "runbook," or
  notes artifact, prefer a live doc.
- Prefer a classic page for knowledge bases, customer-facing help, established
  classic-page spaces, page-only operations, or explicit page requests.
- For non-trivial edits, read current content, save the body locally, edit the
  file, then update with the snapshot token.
- Use the lossless HTML round trip when macros or exact storage representation
  matter.
- Use `--dry-run` only for explicit preview or validation requests, or for
  unusually risky edits where direct execution was not requested.
- Read back the content or space after mutation and report its URL.

## Handoffs

- Load `twg-context-discovery` for related Jira work, projects, goals, or
  dependencies; load `twg-responsibility-routing` for people, ownership,
  authority, or escalation.
- Load `twg-status-rollups` when pages contribute to a broader status report.
- Load `twg-operational-health` for runbooks, incidents, PIRs, or reliability
  evidence.

## References

- `references/content.md` - content types, reads, writes, and exports
- `references/editing.md` - concurrency-safe body editing
- `references/spaces.md` - space lifecycle and hierarchy
- `references/querying.md` - CQL and fuzzy discovery
- `references/body-formats.md` - HTML, markdown, mentions, and special formats

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [atlassian](https://github.com/atlassian)
- **Source:** [atlassian/twg-cli](https://github.com/atlassian/twg-cli)
- **License:** Apache-2.0
- **Homepage:** https://teamwork-graph.atlassian.com/cli/install

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-atlassian-twg-cli-twg-confluence
- Seller: https://agentstack.voostack.com/s/atlassian
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.