Improve tool execution and add measured harness evaluation - #5
Merged
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Bash output limits could terminate otherwise valid commands, leaving the agent to repair interrupted work. This branch separates display truncation from execution, retains bounded overflow logs, and fixes completion handling for inherited output pipes.
It also adds stable Codex session affinity, multiple tool calls per response, validated same-file batch edits, and local evaluation and session-diagnosis tooling. The README now focuses on setup and key benchmark results, with separate usage, architecture, and evaluation guides.
On 14 matching complete-usage runs, updated Ava cost 10.3% less than earlier Ava; the gap against historical Pi narrowed from 14.2% to 2.5%. The full follow-up passed 14/15 with one TLS interruption and unknown request usage. This is a small development comparison, not evidence of general parity.
Validation: 155 Python tests and 10 frontend tests passed, along with frontend build checks, Ruff, mypy, and dependency-lock validation. The final documentation changes passed local link, anchor, benchmark-figure, and diff checks.
Detailed experiment records and private datasets remain local. CI runs code checks and unit tests, not task evaluations.