Skip to content

Add Space-level Prompt and Skill Frames - #161

Open
Yuqing (mydmdm) wants to merge 2 commits into
mainfrom
fix/issue-160
Open

Add Space-level Prompt and Skill Frames#161
Yuqing (mydmdm) wants to merge 2 commits into
mainfrom
fix/issue-160

Conversation

@mydmdm

@mydmdm Yuqing (mydmdm) commented Sep 4, 2026

Copy link
Copy Markdown
Contributor

Summary

  • recognize explicitly authored prompt / prompt: ... Frames as Space-level instructions for fixed Agent Nodes
  • recognize skill / skill: ... Frames as live modules appended to authenticated GET /skill responses
  • share deterministic Text/Note rendering, diagnostics, provenance filtering, and separate 16 KiB budgets across both Frame channels
  • persist Prompt Frame output at first Agent Node realization while resolving Skill Frames live on each authenticated guide request
  • show zoom-invariant Prompt and Skill pills beside recognized Frame labels using the same shared classifier
  • preserve anonymous bundled-guide behavior and legacy Space skill.md overrides

Validation

  • pnpm typecheck
  • pnpm format
  • pnpm lint:fix
  • focused server, shared classifier, and Web badge regression suites

Closes #160

Yuqing (mydmdm) and others added 2 commits September 4, 2026 15:57
Compile explicitly authored prompt Frames into bounded, durable instructions for fixed Agent Nodes across built-in and ACP runtimes.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Extend authenticated Space guides with live Skill Frame modules and surface Prompt and Skill semantics through shared, zoom-invariant Frame badges.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@mydmdm Yuqing (mydmdm) changed the title Add Space-level Prompt Frames for Agent Nodes Add Space-level Prompt and Skill Frames Sep 4, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Add Space-level prompt Frames for global agent instructions

1 participant