Commit History
-
Keep inference runner exit code stable
Verified signature8b2a653 -
Redesign env as incident operations workflow
Verified signature239f219 -
Add LLM-backed grading and planning
Verified signaturefeaba0b -
Harden env isolation and proxy validation
Verified signaturee65a128 -
Align inference with submission env contract
Verified signature90e57c3 -
Require proxy env vars for submission inference
Verified signature2355508 -
Use API_KEY for LiteLLM proxy calls
Verified signature70e2df8 -
Fix reset endpoint and Python 3.14 compatibility
Verified signatured2739ca -
Finalize submission checks and long inference output format
Verified signaturefea495a -
Switch UI to minimal monochrome dark mode
Verified signatureb92cb08 -
Fix prompt diff highlighting alignment
Verified signature4902ac6 -
Add prompt diff view with paste and copy controls
Verified signature8af76a5 -
Lightly optimize short prompts without mangling words
Verified signature54935d0 -
Preserve whole words and short prompts in optimizer
Verified signature2874632 -
Rewrite README for final submission contract
Verified signaturedee59b7 -
Align inference output with strict hackathon contract
Verified signature5b09e23 -
Add grounded optimization modes and wire dedicated optimizer module
Verified signature64b4475 -
Improve citation guidance, preserve prompt structure, harden baseline, and validate inference
Verified signaturea2071ee -
Fix grounding truthfulness, baseline fallback honesty, docs, and validator coverage
Verified signature4417d40 -
Enable CPU PyTorch in Space container for context tuning
Verified signature93ba7d9 -
Add PyTorch context tuning with demo-seeded retrieval and UI grounding stats
Verified signatureaf386b1 -
Add gold citations and multi-pack enterprise corpus switching
Verified signature0badef4 -
Expose grounding and citation signals in frontend and baseline
Verified signature51265a9 -
Add citation-aware grading and swappable corpus path
Verified signaturef071f9a -
Redesign corpus and tasks around support triage and incident response
Verified signature3b45541 -
Finalize single-template UI and enterprise workflow framing
Verified signature90d41df -
Force fresh template reads and add visible build marker
Verified signature9558e85 -
Extract UI template and reframe benchmark as enterprise workflow
Verified signaturea369a91 -
Fix reset default and harden baseline runner semantics
Verified signaturea91b8e7 -
Clean repo artifacts and harden submission assets
Verified signature775814d
-
Use full-height left-right prompt workspace layout
Verified signature7add4b7 -
Restyle UI with cleaner Apple-like design language
Verified signature7ef9dd2 -
Fix stale optimization summary after prompt edits
Verified signature4b36f21 -
Fix optimize summary updates and enforce shorter output
Verified signature5dbc5ac -
Make optimize-prompt actually shorten the prompt
Verified signaturea2ca82e -
Simplify UI to prompt-in optimized-prompt-out flow
Verified signature21899c3 -
Improve custom prompt retrieval relevance and project chunk ranking
Verified signature580e200 -
Show optimized prompt output and real compression effect
Verified signature1af81f4 -
Rebuild UI and harden frontend-backend action flow
Verified signature7607478 -
Allow compressing visible chunks before selection
Verified signaturea4fd1cc -
Use real retrieved chunks for custom prompts
Verified signature8b5fe59 -
Make UI actions auto-start episodes from custom prompt
Verified signature2b2477c -
Fix custom prompt UI and token optimizer flow
Verified signaturef8d0eb2 -
Add interactive root UI and Streamlit app
Verified signaturede095eb -
Initial OpenEnv submission
Verified signature5e1433f -
initial commit
Verified signature96be722