Add 10-min GPU idle-shutdown; make Preview/Code toggle a real button
Claude
Unverified signature
af9f4e6
The simulated NVIDIA GPU badge now sleeps after 10 minutes without a
Run Test, mirroring real rented-GPU cost behavior; the next run pays
an honest, logged cold-start delay to wake it. Also replaced the
easy-to-miss vertical edge tab with a bold fixed pill button
(bottom-right) for Preview/Code, since the original was overlapping
other text and going unnoticed.
Add Preview/Code toggle: real source files from the reference targets
Claude
Unverified signature
0f54cd7
For a pipeline we didn't design and have no connectivity history on,
the README summary in the Project tab isn't enough to actually work on
it. The new CODE tab (fixed to the right edge) fetches real entrypoint
files (package.json, Dockerfile, index.*, etc.) from both reference
targets via the GitHub/HF Space APIs and shows their real content in a
file browser, so surgery or an autopsy on someone else's pipeline is
actually possible instead of guessing from a description.
Add explicit 'End Session' to archive test runs as a documented decision
Claude
Unverified signature
9cf05d5
Runs no longer archive individually the moment they finish — they
accumulate in the current session until End Session is pressed, which
writes one Gallery entry covering the whole session: run count,
pass/blocked breakdown, and a performance summary pulled from the
pipeline's own audit findings and realism-critical stages. This makes
the Gallery a database of documented outcomes rather than a raw log
of every run.
Give Jessica a Play button: real LLM line + real speech, audio-reactive puppet
Claude
Unverified signature
6553b8c
/api/avatar-speak asks the same NIM/Anthropic model already wired up
to generate one short first-person line as the avatar. The browser's
real speech synthesis speaks it, and its actual boundary/playback
events drive a mouth-open pulse and periodic blink overlaid on
Jessica's photo. This is real audio driving real motion — not neural
lip-sync, which would need a GPU-hosted video model instead.
Archive completed Test Bench runs to a paginated Gallery page
Claude
Unverified signature
4df7155
Every run that reaches a terminal state (done/blocked) is retained
with its checksum, byte size, LOC, node count, and full log — browser
localStorage, not a shared database. The Gallery page (grid, newest
first, IntersectionObserver pagination matching the Briefing page)
lets you browse and expand past runs. Gallery added to the top nav.
Add persistent top-center nav across Home/Pipeline/Briefing/Test Bench
Claude
Unverified signature
e79b700
A global nav bar in the root layout replaces the scattered one-off
links that existed only on some pages, so every page is reachable
from every other page. Pipeline and Test Bench's full-viewport layouts
now size against the new nav height instead of assuming 100dvh alone.
Add artifact-immutability lock at the grid boundary
Claude
Unverified signature
e2dea69
Nodes get a SHA-256 checksum + byte-size + LOC snapshot locked in as
they render in the React Flow grid. Before a test run can use that
data it re-verifies against the lock and refuses to proceed on a
mismatch; the lock stays checked-out (with a visible timeout) until
the run returns and re-arms it. Jessica's reference photo is a fixed
UI asset, not a pipeline node, so it never enters the checkout cycle.
Wire Jessica reference photo into Test Bench Default Model slot
Claude
Unverified signature
04e34a2
Tracked via Git LFS to satisfy Hugging Face's binary-file push policy.
Add Test Bench: grid-inspect pipeline stages against real reference implementations
Claude
Unverified signature
e1e33d1
Grid layout of compact stage cards next to an Artifact panel with Project
(live-fetched HF Space + GitHub context), Console (scripted, honestly-labeled
simulated GPU test run under a 3-min budget), and Samantha tabs. DEEP DIVE
seeds Samantha with a compare-against-reference question. Extracted repo/space
fetch helpers into src/lib/repo-fetch.ts for reuse with generate-pipeline.
