ndif skills: corrections from two fresh-user verification runs - #11
Merged
Merged
Conversation
A Docker Hub run and a PyPI run of the 0.1.0 release, each done by an agent with only the public artifacts and these skills. Both succeeded; both left corrections: the jq recipe for model keys did not run in the image and listed COLD stubs; a fresh server's hundreds of COLD entries are the HF cache listing; route 1 had no shutdown guidance; route 3 assumed a checkout and its conda line stops on the anaconda ToS prompt; Ray boots in 15 s, not 60-90, and a dead Ray looks the same as a slow one; NDIF_RAY_TEMP_DIR has a 107-byte socket-path limit; ndif queue omits a pre-deployed model until its first request; schedule.end_time is an eviction window, not a teardown; the telemetry fail-open claim was false on 0.1.0; and a too-long prompt dies as a CUDA device-side assert. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Two agents ran the 0.1.0 release as new users — one from the Docker Hub image, one from PyPI with no Docker — using only the public artifacts and these skills. Both completed the scenario (server up, gpt2 trace, >5 MB result, pre-deploy a second model, inspect queue/status, shut down). This folds in every correction they reported; the code-level ones are ndif-team/ndif#287 and a from-source fixes PR.
🤖 Generated with Claude Code