feat(workspaces): inbox hand-off + neutral instruction.md source#248
Merged
Conversation
… inbox The chat workspace already exposes inbox_push (per-workspace MCP), but nothing told the agent to use it — so the Inbox stayed empty. Add a line to the chat instruction: when you finish something the user should see (a shortlist, thesis, rotation snapshot, a decision), push it to their inbox with the file(s) + a short note. Activates the one-way workspace→user collaboration loop with no new code — a prototype. The inbox renders the docs live, so pushed dossier files keep updating as the agent CRUDs them. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
…E.md + AGENTS.md Claude Code wants CLAUDE.md, Codex wants AGENTS.md — a standards turf war we don't need to pick a side in. The launcher already wrote one composed result to both filenames; this renames the single template SOURCE from CLAUDE.md to the neutral instruction.md, so the source isn't named after either CLI. The injector reads <template>/instruction.md and writes the composed (persona + instruction) bytes to both CLAUDE.md and AGENTS.md. A third CLI = one more copy target, no new source, no sync risk (there's still only one source). Also hardened the injector golden to mirror the persona precedence (data/brain/persona.md override → shipped default) so a live persona override doesn't break the byte assertion. tsc clean; full suite 1718; creation e2e green. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
luokerenx4
changed the title
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
A collaboration prototype, no new code. The chat workspace already exposes the
inbox_pushtool (per-workspace MCP), but nothing instructed the agent to use it — so the Inbox stayed empty. One line in the chat instruction now tells the agent: when you finish something the user should see (a shortlist, a thesis, a rotation snapshot, a decision), push it to their inbox with the file(s) + a short note.This activates the one-way workspace→user loop. Nicely,
inbox_pushrendersdocslive (no snapshot), so pushed dossier files keep reflecting the agent's edits across sessions.Test plan
Boundary touch
None — chat template instruction content only.
🤖 Generated with Claude Code