Auto-start CPU sandboxes for sessions (#200)

Steer agent to HF kernels instead of pip install flash-attn (#204)

Make sandboxes private (#195)

feat(telemetry): track 5 untracked Bedrock call sites for full cost attribution (#179)

Track Pro conversions + credits top-up; emit user_id + new KPIs (#174)

Bill HF Jobs by namespace credits, drop Pro and org-join gating (#172)

fix(sandbox): retry cleanup on transient failures + add backstop sweeper (#163)

Embed trackio dashboard in chat for jobs and sandboxes (#161)

Add web search so the agent can cite current sources (#159)

Add Slack gateway (#116)

Protect sandbox control-plane routes (#142)

Keep repetition guard internal (#144)

Run jobs under user or org accounts with upgrade / org-pick UX (#132)

Fix CLI rendering corruption and split CLI/frontend model defaults (#121)

feat(observability)

feat: route Claude via Bedrock inference profile (#82)

feat: enable Anthropic prompt caching on system prompt and tools (#69)

fix /model switcher with effort params

Fix parallel-research display: per-call stats in CLI and web UI (#49)

Strip provider_specific_fields from research sub-agent messages

Route HF inference through /v1 auto-router + add reasoning_effort knob

Fall back to user HF token for router when INFERENCE_TOKEN is unset (#39)

Rename CLI entry and Space assets from ml-agent to ml-intern

Rewrite research workflow to lead with literature mining

Fix sandbox HF_TOKEN not available inside container

Minimize S2 API calls: drop paper_details enrichment, add response cache, rate limit only with API key

Add Semantic Scholar integration: citation graph, filtered search, snippet search, recommendations

Use sandbox read endpoint for script resolution to avoid bash truncation

Remove outdated-API flagging line from research prompt

Add SOTA-awareness to research sub-agent system prompt

Improve agent loop, tools, and research subagent for long-running autonomous tasks

Use Haiku 4.5 for research sub-agent

Use Haiku for research sub-agent when main model is Anthropic

Research stats tick every second via background timer

Live research stats: tool count, tokens, elapsed time in header

Revert "Sub-agent log: tool_name → output_preview on one line"

Sub-agent log: tool_name → output_preview on one line

Log 20-char output preview for sub-agent tool calls

feat: add headless CLI mode with local filesystem tools and rich terminal rendering

Deduplicate file re-reads via content hashing for local and sandbox tools

fix: move import/signature validation to sandbox, keep host-side lightweight

feat: emit tool_log events from research sub-agent for UI liveness

feat: add research sub-agent tool, slim down main agent system prompt

Improve tool output quality: ANSI stripping, tail-biased truncation, noise suppression

feat: cancel sandbox creation on user interrupt

feat: kill sandbox processes and cancel HF jobs on user interrupt

fix: retry sandbox API calls on transient failures

chore: double sandbox command timeouts

chore: increase sandbox sleep_time to 45 min

feat: CLI local mode, slash commands, interrupt support; remove lmnr; frontend fixes