4fb0059556
* fix(rag): per-index chunk floors — journals and learnings were never indexed The global 200-char garbage floor (sized for code/doc chunks) discarded every templated journal note and most distilled org-memory lessons, silently: ingest returned success with zero chunks, so agent journals and learnings were never retrievable via RAG. IndexConfig now carries a per-type min_chunk_length (journals 40, learnings 80, others unchanged). * fix(mcp): git-readonly tools default project_slug from the container env Agents 404ed /api/git/status with 'Project not found: roboco' — the tools made the LLM supply the slug and six doc examples taught a slug that matches no registered project. The tools now fall back to the ROBOCO_PROJECT_SLUG the orchestrator already injects, and the stale examples are corrected. * feat(rag): startup backfill re-ingests zero-chunk journals and learnings Before the per-index chunk-floor fix, ingest() returned success with chunk_count=0 for undersized content: every historical journal entry and distilled learning below the (then-global) 200-char floor was durably recorded in journal_entries but silently never got a chunks_journals / chunks_learnings row, and no exception meant the existing dead-letter (rag_index_failures) never saw it either. Extends the startup reconcile (roboco/api/app.py _reconcile_rag_indexes) with a new pass: backfill_unindexed_journals (roboco/services/ rag_index_failures.py) queries journal_entries for rows missing from each vector table and re-ingests them through the same live code paths (_reindex_journal_entry / record_learning). Journals and learnings are backfilled independently since a LEARNING entry can clear the (lower) JOURNALS floor while still failing the (higher) LEARNINGS floor — a learning's doc_source is a content hash, not the entry id, so presence there is checked by hashing each candidate the same way LearningsIndexPlugin.record_learning does and batch-querying chunks_learnings for those exact sources. Bounded to 200 rows per pass per boot (converges over restarts on a larger backlog) and best-effort per row (one failure never aborts the pass). Rows still under the current floor are excluded by a length filter in the SELECT so they are never retried forever, and private entries are excluded from the JOURNALS pass exactly like the live indexing path. * test(rag): scope backfill assertions to their own rows --------- Co-authored-by: Renn F <rennf93@users.noreply.github.com>
2.4 KiB
Project & Workspace Tools
Overview
There is no roboco_project_* or roboco_workspace_* agent tool. Agents do not create projects, manage git tokens, or ensure workspaces. Those are handled for you:
- Workspaces are auto-cloned by the orchestrator (
WorkspaceService). Your per-agent clone of the project repo is created the first time you claim work on it — you never call a workspace tool. Branches are auto-created oni_will_work_on()/claim_review(); you don't rungit checkouteither. - Project registration and git-token management are operator actions done through the control panel / HTTP API, not from inside an agent container. Tokens are encrypted at rest; the agent container never sees the PAT (it is injected into git operations server-side and scrubbed from URLs).
What a task already tells you
A task carries its project linkage; you don't look it up with a tool. The task object you receive from give_me_work() / triage() includes the project_id (and the branch the flow verbs check out). Acceptance criteria and the project context come back inline on the Envelope.
Inspecting the repo
Read-only git inspection is available through the roboco-git-readonly MCP server (developers and QA):
# project_slug is optional on all four — omit it and your own project
# is used (from this agent's environment).
roboco_git_status()
roboco_git_log()
roboco_git_diff()
roboco_git_branch_list()
There is no roboco_git_commit / _push / _checkout / _create_pr / _merge_pr tool. Commits go through the commit content tool (auto- prefixed with [task-id], auto-pushed by the choreographer); PRs open at open_pr time; merges are a PM complete operation.
Finding project knowledge
To learn how a project's codebase is laid out or how a subsystem works, query the knowledge base rather than a project tool:
roboco_kb_search(query="rate limiting redis", project="roboco-api",
index_types=["code", "documentation"])
roboco_ask_mentor(question="How is auth wired up in this project?")
PM note: creating work
PMs create work with the delegate flow verb (a subtask under the current parent task), not a project/task-create tool. delegate takes an optional project_id; the parent task's project is inherited when you omit it. There is no agent-facing standalone project- or task-create tool.