Skip to content

title: "## studyio — hypothesis-aware research orchestration layer for projio date: 2026-04-07 timestamp: 20260407-225436-752515 tags: [idea] source: agent-observation project_primary: projio capture_id: 20260407-225434-7671b4 confidence: 1.0 status: done transcript_file: /storage2/arash/worklog/workflow/captures/20260407-225434-7671b4/transcript.txt

studyio — hypothesis-aware research orchestration layer for projio

Problem

Projio's subsystems (pipeio, notio, biblio, manuscripto, codio, indexio) cover the "hands and memory" of agentic research — pipeline execution, note capture, literature management, manuscript assembly, code documentation, search. But there is no subsystem that understands the scientific reasoning layer: hypotheses, evidence, milestones, and the feedback loop between plan and results.

Today an agent can run a pipeline, capture a note, and build a manuscript section. But it cannot answer: "What pipeline runs are needed to test H3?" or "Is there enough evidence to write the Results section for H4?" or "What should the supervisor see this week?"

Research workflow ontology

   PLAN ──→ EXECUTE ──→ OBSERVE ──→ INTERPRET ──→ WRITE
     ↑                                              │
     └──────────────────────────────────────────────┘
Phase Folder Projio subsystem Entities
Plan plan/ worklog (goals) hypotheses, milestones, roadmap
Execute code/pipelines/ pipeio flows, mods, rules, configs
Observe log/ notio observations, results, decisions
Interpret log/ + bib/ notio + biblio result notes + literature
Write manuscript/ manuscripto sections, figures, citations

Coverage assessment (~60% today)

Strong: pipeline execution (pipeio), literature (biblio), manuscript (manuscripto), code docs (codio), notes (notio), cross-project goals (worklog).

Missing (~40%): hypothesis→evidence reasoning, evidence→manuscript linking, progress→supervisor reporting, hypothesis-aware pipeline dispatch.

Proposed MCP tools

Planning: study_hypotheses, study_milestone_update, study_progress, study_next, study_roadmap_refresh

Evidence: study_result_create (structured result notes with hypothesis/milestone/metric fields), study_evidence_collect, study_evidence_gap

Supervisor: study_report (progress summary), study_meeting_prep

Bridge: study_dispatch (hypothesis→pipeline runs), study_pipeline_impact

Integration

                studyio
               /   |   \
        worklog  pipeio  notio
        (goals)  (runs)  (notes)
           |       |       |
           v       v       v
        milestones flows  results
                \   |   /
                 v  v  v
               manuscripto

Implementation phases

  1. Data model — formalize hypotheses.yml, structured result note frontmatter. Validate with pixecog.
  2. Read-only tools — study_hypotheses, study_progress, study_evidence_gap. Parse existing markdown/YAML.
  3. Write tools — study_result_create, study_milestone_update. Agent records results.
  4. Orchestration — study_next, study_dispatch, study_report. Agent plans and executes.
  5. Manuscript integration — evidence_collect feeds manuscript_section_context. Agent drafts when ready.

Full detail

Full design with entity relationship diagrams, agentic workflow example, and worklog/studyio relationship is in pixecog note idea-arash-20260407-225257-089158.

Source context: pixecog

PixEcog (pixecog): Neuropixels and ECoG dataset and analysis

Recent commits: 

9b2f6fa Scaffold ecephys TTL removal mod, flow overview + mod docs, demo notebook
80194af Add TTL characterization & removal demo notebook (preprocess_ieeg)
dc93496 Update mkdocs pipeline nav

README:

type: readme

Quick Start for Collaborators

Follow this checklist to get started with Pixecog documentation and workflows.

🐀 Pixecog Project — Compact Overview

Core principles

  • One immutable BIDS raw dataset (raw/) as the canonical baseline
  • Each analysis pipeline ha
  • idea-arash-20260403-172004-817050.md — Directly related: proposes new projio ecosystem skills/subsystems; studyio would add a hypothesis-reasoning layer alongside those candidates
  • deep-research-pipeio-scope.md — Parallel scoping exercise — what a projio subsystem should absorb, reuse, or build; same design question studyio raises for the scientific reasoning layer
  • idea-arash-20260330-174518-164647.md — pipeio v2 roadmap defines the ecosystem studyio must integrate with; studyio sits atop pipeio's execute layer
  • idea-arash-20260407-171834-423514.md — Configurable docs paths for subsystem-owned docs — directly relevant if studyio becomes a first-class projio subsystem with its own docs namespace
  • idea-arash-20260407-010140-074558.md — Auto-generated CHANGELOG from task/datalad completions touches the same milestone-tracking gap studyio aims to close