v0.1.6

Highlights

Schedule harness, active learning upgrades, skills/KB tooling, and Darkroom UI polish since v0.1.5.

Scheduler

• Interval, 5-field cron, and one-shot schedules (every 1h, 0 9 * * *, 30m, ISO time)
• Claim-before-fire (no double-run), pause/resume, run-now, repeat limits
• Agent schedule tool: create / list / update / pause / resume / remove / run / runs
• API: POST /api/schedules/{id}/run|pause|resume plus schedule field on create
• Scheduler UI: schedule string input, Run / Pause actions

Active learning

• Package-style learning harness (state, digest, prompts, runner)
• Sticky memory/skill nudges (cooldown no longer burns a trigger)
• Cumulative skill tool-iters across turns; refine when a skill was touched
• Structured review digest + optional cheaper review model profile
• Task-local isolation so reviews don’t block other agents’ counters

Skills & knowledge

• Skill catalog search and SKILL.md body rendering
• Bulk skill assign on agents (filter + select-all)
• Knowledge base upload for PDF / DOCX / text

UI

• Darkroom workbench for agent studio and workplaces
• Darkroom chat chrome cleanup
• Gear icon for System nav

Other

• web_search uses DuckDuckGo HTML results

Install

App (managed):

curl -fsSL https://raw.githubusercontent.com/Alg0rix/tomo/v0.1.6/scripts/install.sh | bash

Connector:

curl -fsSL https://raw.githubusercontent.com/Alg0rix/tomo/v0.1.6/scripts/install-connector.sh | bash
# pin: TOMO_CONNECTOR_VERSION=v0.1.6

Wheel: tomo-0.1.6-py3-none-any.whl (attached)

What's Changed

• feat: schedule harness and learning distill upgrades
• fix(ui): Darkroom workbench for agent studio and workplaces
• feat(agents): bulk skill assign with filter and select-all
• feat(skills): search catalog and render SKILL.md body
• fix(ui): Darkroom chat chrome, de-slop clones, KB search
• fix(web_search): use DuckDuckGo HTML results
• feat(kb): upload PDF/DOCX/text into knowledge base
• fix(ui): use gear icon for System nav
• chore: bump version to 0.1.6

Full Changelog: v0.1.5...v0.1.6