Repository navigation
Chore/workflows - #2836
Chore/workflows#2836tommywagz wants to merge 42 commits into
Conversation
Co-authored-by: Shahin Saadati <3443249+happyhuman@users.noreply.github.com>
|
This PR changes 23783 lines across 106 files. GitHub's diff API will not serve a diff above 20000 lines, so automated AI review was skipped. Splitting it up would get each part reviewed. |
|
This PR changes 23783 lines across 106 files. GitHub's diff API will not serve a diff above 20000 lines, so automated AI review was skipped. Splitting it up would get each part reviewed. |
github-actions
Bot
added
the
status/not-ready-for-review
|
Ready for review. All checks pass. Re-requested review from |
|
This PR changes 23778 lines across 106 files. GitHub's diff API will not serve a diff above 20000 lines, so automated AI review was skipped. Splitting it up would get each part reviewed. |
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
There was a problem hiding this comment.
Automated House Rules review — 0 finding(s).
Also, on lines this PR does not change:
plugins/README.md:1— [MINOR] this PR changes 3 file(s) under .agents/skills/ alongside 4 recipe file(s). Repo skills and recipes are separate concerns and belong in separate PRs (AGENTS.md:14)
| description: 'agy model id' | ||
| required: false | ||
| type: 'string' | ||
| default: 'gemini-3.5-flash' |
There was a problem hiding this comment.
[MINOR] Is there a reason this defaults to gemini-3.5-flash instead of gemini-3.7-flash? AGENTS.md:40 specifies gemini-3.7-flash as the default.
| if first_brace != -1 and last_brace != -1 and last_brace > first_brace: | ||
| snippet = raw[first_brace : last_brace + 1].strip() | ||
| try: | ||
| data = json.loads(snippet) |
There was a problem hiding this comment.
[MINOR] Can we extract the repeated json.loads and dictionary check into a helper? The same try/except block is duplicated three times in this function.
|
|
||
| # When allow_comment is False (initial routing), or when Option 4 is chosen, | ||
| # resolve assignee based on path routing: | ||
| if option == Option.ACKNOWLEDGE_AND_ASSIGN or not allow_comment: |
There was a problem hiding this comment.
[MINOR] Can we combine these branches? It repeats the identical if option == Option.ACKNOWLEDGE_AND_ASSIGN or not allow_comment: check evaluated just above.
|
|
||
| if not args.result.exists(): | ||
| print(f"Error: Result file not found: {args.result}", file=sys.stderr) | ||
| return 2 |
There was a problem hiding this comment.
[MINOR] bare status code literal
| except json.JSONDecodeError: | ||
| labels = [x.strip() for x in stripped.split(",") if x.strip()] | ||
| else: | ||
| labels = [x.strip() for x in stripped.split(",") if x.strip()] |
There was a problem hiding this comment.
[MINOR] duplicated fallback splitting logic
|
There is a problem with the Antigravity CLI Hygiene PR review. Please check the action logs for details. |
|
There is a problem with the Antigravity CLI Correctness PR review. Please check the action logs for details. |
…ariable interpolation, and correct template signal typo
…proposals and use file-based catalog prompt
There was a problem hiding this comment.
Automated House Rules review — 1 finding(s).
Also, on lines this PR does not change:
core/python/long-horizon-harness/.gitignore:1— [MINOR] this PR changes 3 file(s) under .agents/skills/ alongside 42 recipe file(s). Repo skills and recipes are separate concerns and belong in separate PRs (AGENTS.md:14)
| @@ -0,0 +1,95 @@ | |||
| """Fail a graded eval run whose cases scored below the threshold. | |||
There was a problem hiding this comment.
[MINOR] no licence header on this file; 5 .py file(s) have none while 449 carry the full block (core/python/long-horizon-harness/scripts/probe_context_cache.py; and 4 more)
There was a problem hiding this comment.
Automated Maintainability review — 1 finding(s).
core/python/long-horizon-harness/AGENTS.mdcore/python/long-horizon-harness/tests/eval/datasets/clarify_gate-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/compression_quality-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/guardrail_halt-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/safety-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/skill_curation-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/slash_commands_and_reload-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/tool_selection_core-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/web_search_grounding-dataset.jsoncore/python/long-horizon-harness/tests/eval/evalsets/README.md
Splitting the PR up would get them reviewed.
Round 2 · 5 of this PR's 25 automated comments used · this round is capped at 3 across all reviewers · after this round only Security and Correctness run.
| ) | ||
|
|
||
| # Or if body contains at least 3 distinct template questions/sections even if user altered title: | ||
| if len(matched_signals) >= 3: |
There was a problem hiding this comment.
[MINOR] Can we extract 3 into a named constant like MIN_TEMPLATE_SIGNALS? It serves as the threshold for matching template signals without a title marker.
There was a problem hiding this comment.
Automated Security review — 0 finding(s).
core/python/long-horizon-harness/AGENTS.mdcore/python/long-horizon-harness/tests/eval/datasets/clarify_gate-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/compression_quality-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/guardrail_halt-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/safety-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/skill_curation-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/slash_commands_and_reload-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/tool_selection_core-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/web_search_grounding-dataset.jsoncore/python/long-horizon-harness/tests/eval/evalsets/README.md
Splitting the PR up would get them reviewed.
Round 2 · 5 of this PR's 25 automated comments used · this round is capped at 3 across all reviewers · after this round only Security and Correctness run.
There was a problem hiding this comment.
Automated Correctness review — 0 finding(s).
core/python/long-horizon-harness/AGENTS.mdcore/python/long-horizon-harness/tests/eval/datasets/clarify_gate-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/compression_quality-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/guardrail_halt-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/safety-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/skill_curation-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/slash_commands_and_reload-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/tool_selection_core-dataset.jsoncore/python/long-horizon-harness/tests/eval/datasets/web_search_grounding-dataset.jsoncore/python/long-horizon-harness/tests/eval/evalsets/README.md
Splitting the PR up would get them reviewed.
Round 2 · 5 of this PR's 25 automated comments used · this round is capped at 3 across all reviewers · after this round only Security and Correctness run.
| @@ -0,0 +1,154 @@ | |||
| #!/usr/bin/env python3 | |||
There was a problem hiding this comment.
[MINOR] this PR changes 3 file(s) under .agents/skills/ alongside 1 recipe file(s). Repo skills and recipes are separate concerns and belong in separate PRs (AGENTS.md:14)
github-actions
Bot
added
the
status/not-ready-for-review
github-actions
Bot
removed
the
status/not-ready-for-review
Updated the github workflows to not immediately comment on new issues with long sycophantic messages, instead opting to just route the issue to the appropriate maintainer. The messages often take the liberty of answering issues with potentially misleading information to the contributors without maintainer approval.