Mission Note Mentions
Share notes across projects and @-mention them inside an Argus mission note — the mentioned content expands inline, read-only, when the manager reads its mission.
An Argus manager reads its instructions from a mission note. Mission Note Mentions let you compose that mission out of other Scape objects: @-mention a note and its content is pulled in inline when the manager reads the mission; @-mention a repo, table, playbook, chart, integration, link, or chat room and it resolves to a one-line pointer the agent can act on.
This makes missions modular — write one shared "Standards" note and mention it from every mission, instead of copying the same instructions into each one.
Sharing a note across projects
Notes are scoped to a project. To let a note be mentioned from a mission in any project, turn on Share Across Projects:
- From the notes list — right-click a note and toggle Share Across Projects.
- From an open note — open the note-tab ⋯ menu and toggle Share Across Projects.
A checkmark appears next to the toggle when it's on, and a small people icon appears in the note's header (hover it for a "Shared across projects" tooltip).
When you @-mention inside a note, the picker offers your current project's objects plus every shared note, so a shared note is reachable from anywhere.
Mentioning a note in a mission
Type @ in a mission note and pick a note. When the manager reads its mission (via the get_note MCP tool), each mentioned note is expanded in full, inline, under a dedicated section — the manager sees the mission and the mentioned content in one read.
A note is expanded when it is shared across projects, or when it lives in the same project as the mission note. A mentioned note that is neither renders as a short placeholder instead of its content, so nothing leaks across a project boundary you didn't open.
Limits and safety
Expansion is bounded so a mission can't pull in an unbounded amount of text:
| Limit | Value | Behavior when exceeded |
|---|---|---|
| Depth | 2 levels | A mention nested deeper than 2 renders as a placeholder instead of expanding. |
| Size | 64 KB per note body | The body is truncated with a --- expansion truncated (size limit) --- marker. |
Cycles are handled. If note A mentions B and B mentions A, A (the mission) is already being read, so B expands once and its mention of A renders as "refers back to this note." A note mentioned twice expands the first time and shows "already expanded above" the second — there is no infinite loop.
Mentioning other objects in a mission
Mention something other than a note and it resolves to a one-line pointer — the human-readable name plus the exact id or path an agent needs to reach it in a single tool call:
| Chip | Resolves to |
|---|---|
| Repo | The repo's local path (or remote URL) — ready for create_worktree / create_session. |
| Table | The store id — for describe_data_store / query_data_store. |
| Playbook | The playbook id — for get_playbook / run_playbook. |
| Chart | The chart id — for get_chart. |
| Integration | The integration type and id — for get_integration_health / diagnose_integration. |
| Link | The URL — for open_url. |
| Chat room | The room id — for post_backchannel / list_backchannels. |
Unlike shared notes, these object pointers never cross a project boundary — a chip resolves only when it belongs to the same project as the note that contains it. Even inside a shared note, its object chips stay scoped to that note's own project, so a mission can never be handed a working directory, table, or room from a project you didn't intend.