Repository navigation
feat: a task-scoped default repo for jobs bound to one repo; release v0.12.0 (#72) - #73
Conversation
… to one repo; release v0.12.0 (#72) `resolve_repo` now resolves explicit repo > the task's scoped default > the configured default > env. `gh_issue.scoped_default_repo(repo)` is a context manager over a ContextVar, so a host plugin can make the repo of a job it runs (the pr-reviewer's review round) the default for every tool call in that job and the tasks it starts, and nowhere else. An explicit repo still wins; a malformed or empty scope changes nothing. Why: Vera reviews 170 repos with default_repo "", so a finder that omitted `repo` read github.repos[0] (protoAgent) while reviewing mythxengine#966 (pr-reviewer-plugin#367). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
There was a problem hiding this comment.
QA panel review — PASS
code-review-structural · head 8f42d562086b · formal
Low-risk change: adds a task-scoped default repo via a ContextVar context manager with solid test coverage. The one thing to fix first is github_status at read_tools.py:846, which still reports the agent-wide default and will mislead the model inside a scoped job. Verification confirmed both surviving findings and refuted the README/uv-drift nit (CI actually uses pip, not uv, so the README is correct). No panel disagreement; cross-file and conventions lanes came back clean. No coverage gaps.
Findings
| Severity | Location | Finding | Verified | |
|---|---|---|---|---|
| 🟡 | minor | read_tools.py:846 |
github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the m… | confirmed |
| 🟡 | minor | pyproject.toml:1 |
pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = "auto"), making uv run pytest fail in a clean environment. | confirmed · nearby, not gating |
findings JSON (machine-readable)
[
{
"file": "read_tools.py",
"line": 846,
"severity": "minor",
"category": "correctness",
"claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).",
"evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).",
"raw_ids": [
"find_correctness:9ecb5ed0ed21acf3e8f2e39f:1",
"find_removed_behavior:43f62a729df77e4f2a17852c:1"
],
"verdict": "confirmed",
"note": "read_tools.py:846 reads `current_default(default_repo)` directly; search_pr_repo confirms `scoped_default` appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts `scoped_default()` between explicit and `current_default`, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."
},
{
"file": "pyproject.toml",
"line": 1,
"severity": "minor",
"category": "build-release",
"claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.",
"evidence": "The [project] table (lines 1\u20135) has no `dependencies` or `optional-dependencies` key. Line 16 sets `asyncio_mode = \"auto\"`, which requires the `pytest-asyncio` plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is `uv run pytest`.",
"source": "protopatch",
"raw_ids": [
"find_structural:1a47a9cf93ac76ff3b679361:1"
],
"verdict": "confirmed",
"note": "pyproject.toml at head has no `dependencies` key and sets `asyncio_mode = \"auto\"` (requires pytest-asyncio). However, the project's actual CI (.github/workflows/ci.yml) uses `pip install -r requirements-dev.txt` + bare `pytest -q`, not `uv run pytest`; the practical impact is limited to anyone who tries the uv workflow. \u2014 nearby: in code this PR did not change (outside its changed lines and the functions they sit in) \u2014 reported, not gated (#232)",
"nearby": true
}
]1 finding(s) excluded from the verdict by in-diff confinement (file not among this PR's changed paths):
read_tools.py(minor) — github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the mo
1 structural finding(s) are nearby notes, not part of the verdict: they sit in code this PR did not change (outside its changed lines and the functions those lines are in). Worth a look; not a request for this PR.
pyproject.toml:16(minor) — pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = "auto"), making uv run pytest fail in a clean environment.
Runner context: observed. Native parent-conversation exclusion and separate history exclusion are checked per execution. Intended prior findings may remain in the rendered prompt. Instruction-source observations describe the assembled system prompt; mandatory repository guidance remains applicable. External instruction sources and runtime additions remain unknown. Tool observations describe bound tools, inherited fence and enforcement; external policy remains unknown. Construction evidence does not establish semantic review coverage.
Raw finding accounting: complete
{"supported": true, "binding": {"attempt": "0e271eceace84f4e81c76db64dcbf9d4", "repo": "protoLabsAI/github-plugin", "head": "8f42d562086b6e9a75e74740f0da849dd2b56f11", "pr": "73"}, "raw": {"find_correctness:9ecb5ed0ed21acf3e8f2e39f:1": {"lane": "find_correctness", "finding": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default repo via current_default(), ignoring the new task-scoped default. Inside a scoped job it tells the model the tools default to a repo they no longer default to.", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo)."}}, "find_removed_behavior:43f62a729df77e4f2a17852c:1": {"lane": "find_removed_behavior", "finding": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "cross-file", "claim": "github_status still reports only the configured default, so inside a scoped_default_repo job it names a different repo than the one the other github_* tools now target when \u0060repo\u0060 is omitted.", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 bypasses the new \u0060scoped_default()\u0060 step in resolve_repo, while the tool's docstring says it reports \"which repo the github_* tools default to\". A finder inside a pr-reviewer round calling github_status to debug a not-found error is told the agent-wide default (e.g. protoAgent) even though omitted-\u0060repo\u0060 calls now hit the job's repo."}}, "find_structural:1904e467e1208b0e846f2c3f:1": {"lane": "find_structural", "finding": {"file": "pyproject.toml", "line": 1, "severity": "nit", "category": "maintainability", "claim": "README instructs pip-based dev setup that conflicts with the uv-based project configuration", "evidence": "README lines 94\u201395 direct developers to \u0060pip install -r requirements-dev.txt\u0060 and then run \u0060ruff\u0060/\u0060pytest\u0060 bare, while the detected project commands all use \u0060uv run\u0060 (e.g. \u0060uv run ruff check .\u0060, \u0060uv run pytest\u0060). A developer following the README would install into the system or a pip venv and run tools without \u0060uv run\u0060, which bypasses the project's managed environment and may hit version mismatche Fix: Update the Develop section to reflect the uv workflow, e.g. \u0060uv sync\u0060 (or \u0060uv run --with ruff,pytest ruff check . && uv run pytest\u0060), and remove or replace the \u0060pip install -r requirements-dev.txt\u0060 in (protopatch confidence: high)", "source": "protopatch"}}, "find_structural:1a47a9cf93ac76ff3b679361:1": {"lane": "find_structural", "finding": {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio, making uv run pytest fail in a clean environment", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060. In a clean checkout, \u0060uv run\u0060 creates an environment resolved from pyproject.tom Fix: Add a \u0060dependencies\u0060 (or \u0060[project.optional-dependencies] dev\u0060) section to pyproject.toml listing at minimum \u0060pytest\u0060 and \u0060pytest-asyncio\u0060, so that \u0060uv sync\u0060 / \u0060uv run\u0060 can build a working test enviro (protopatch confidence: medium)", "source": "protopatch"}}}, "lanes": {"find_correctness": {"state": "observed", "count": 1}, "find_removed_behavior": {"state": "observed", "count": 1}, "find_crossfile": {"state": "observed", "count": 0}, "find_conventions": {"state": "observed", "count": 0}, "find_structural": {"state": "observed", "count": 2}}, "gaps": [], "transitions": [{"stage": "synthesize", "raw_id": "find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "target": "find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "before": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default repo via current_default(), ignoring the new task-scoped default. Inside a scoped job it tells the model the tools default to a repo they no longer default to.", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo)."}, "after": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"]}}, {"stage": "synthesize", "raw_id": "find_removed_behavior:43f62a729df77e4f2a17852c:1", "target": "find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "before": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "cross-file", "claim": "github_status still reports only the configured default, so inside a scoped_default_repo job it names a different repo than the one the other github_* tools now target when \u0060repo\u0060 is omitted.", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 bypasses the new \u0060scoped_default()\u0060 step in resolve_repo, while the tool's docstring says it reports \"which repo the github_* tools default to\". A finder inside a pr-reviewer round calling github_status to debug a not-found error is told the agent-wide default (e.g. protoAgent) even though omitted-\u0060repo\u0060 calls now hit the job's repo."}, "after": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"]}}, {"stage": "synthesize", "raw_id": "find_structural:1a47a9cf93ac76ff3b679361:1", "target": "find_structural:1a47a9cf93ac76ff3b679361:1", "before": {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio, making uv run pytest fail in a clean environment", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060. In a clean checkout, \u0060uv run\u0060 creates an environment resolved from pyproject.tom Fix: Add a \u0060dependencies\u0060 (or \u0060[project.optional-dependencies] dev\u0060) section to pyproject.toml listing at minimum \u0060pytest\u0060 and \u0060pytest-asyncio\u0060, so that \u0060uv sync\u0060 / \u0060uv run\u0060 can build a working test enviro (protopatch confidence: medium)", "source": "protopatch"}, "after": {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060.", "source": "protopatch", "raw_ids": ["find_structural:1a47a9cf93ac76ff3b679361:1"]}}, {"stage": "synthesize", "raw_id": "find_structural:1904e467e1208b0e846f2c3f:1", "target": "find_structural:1904e467e1208b0e846f2c3f:1", "before": {"file": "pyproject.toml", "line": 1, "severity": "nit", "category": "maintainability", "claim": "README instructs pip-based dev setup that conflicts with the uv-based project configuration", "evidence": "README lines 94\u201395 direct developers to \u0060pip install -r requirements-dev.txt\u0060 and then run \u0060ruff\u0060/\u0060pytest\u0060 bare, while the detected project commands all use \u0060uv run\u0060 (e.g. \u0060uv run ruff check .\u0060, \u0060uv run pytest\u0060). A developer following the README would install into the system or a pip venv and run tools without \u0060uv run\u0060, which bypasses the project's managed environment and may hit version mismatche Fix: Update the Develop section to reflect the uv workflow, e.g. \u0060uv sync\u0060 (or \u0060uv run --with ruff,pytest ruff check . && uv run pytest\u0060), and remove or replace the \u0060pip install -r requirements-dev.txt\u0060 in (protopatch confidence: high)", "source": "protopatch"}, "after": {"file": "pyproject.toml", "line": 1, "severity": "nit", "category": "conventions", "claim": "README instructs pip-based dev setup (pip install -r requirements-dev.txt, bare ruff/pytest) that conflicts with the uv-based project configuration (uv run ruff, uv run pytest).", "evidence": "README lines 94\u201395 direct developers to \u0060pip install -r requirements-dev.txt\u0060 and then run \u0060ruff\u0060/\u0060pytest\u0060 bare, while the detected project commands all use \u0060uv run\u0060 (e.g. \u0060uv run ruff check .\u0060, \u0060uv run pytest\u0060).", "source": "protopatch", "raw_ids": ["find_structural:1904e467e1208b0e846f2c3f:1"]}}, {"stage": "verify", "raw_id": "find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "target": "find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "before": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"]}, "after": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}}, {"stage": "verify", "raw_id": "find_removed_behavior:43f62a729df77e4f2a17852c:1", "target": "find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "before": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"]}, "after": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}}, {"stage": "verify", "raw_id": "find_structural:1a47a9cf93ac76ff3b679361:1", "target": "find_structural:1a47a9cf93ac76ff3b679361:1", "before": {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060.", "source": "protopatch", "raw_ids": ["find_structural:1a47a9cf93ac76ff3b679361:1"]}, "after": {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060.", "source": "protopatch", "raw_ids": ["find_structural:1a47a9cf93ac76ff3b679361:1"], "verdict": "confirmed", "note": "pyproject.toml at head has no \u0060dependencies\u0060 key and sets \u0060asyncio_mode = \"auto\"\u0060 (requires pytest-asyncio). However, the project's actual CI (.github/workflows/ci.yml) uses \u0060pip install -r requirements-dev.txt\u0060 + bare \u0060pytest -q\u0060, not \u0060uv run pytest\u0060; the practical impact is limited to anyone who tries the uv workflow."}}, {"stage": "verify", "raw_id": "find_structural:1904e467e1208b0e846f2c3f:1", "target": "find_structural:1904e467e1208b0e846f2c3f:1", "before": {"file": "pyproject.toml", "line": 1, "severity": "nit", "category": "conventions", "claim": "README instructs pip-based dev setup (pip install -r requirements-dev.txt, bare ruff/pytest) that conflicts with the uv-based project configuration (uv run ruff, uv run pytest).", "evidence": "README lines 94\u201395 direct developers to \u0060pip install -r requirements-dev.txt\u0060 and then run \u0060ruff\u0060/\u0060pytest\u0060 bare, while the detected project commands all use \u0060uv run\u0060 (e.g. \u0060uv run ruff check .\u0060, \u0060uv run pytest\u0060).", "source": "protopatch", "raw_ids": ["find_structural:1904e467e1208b0e846f2c3f:1"]}, "after": {"file": "pyproject.toml", "line": 1, "severity": "nit", "category": "conventions", "claim": "README instructs pip-based dev setup (pip install -r requirements-dev.txt, bare ruff/pytest) that conflicts with the uv-based project configuration (uv run ruff, uv run pytest).", "evidence": "README lines 94\u201395 direct developers to \u0060pip install -r requirements-dev.txt\u0060 and then run \u0060ruff\u0060/\u0060pytest\u0060 bare, while the detected project commands all use \u0060uv run\u0060 (e.g. \u0060uv run ruff check .\u0060, \u0060uv run pytest\u0060).", "source": "protopatch", "raw_ids": ["find_structural:1904e467e1208b0e846f2c3f:1"], "verdict": "refuted", "note": "The CI workflow (.github/workflows/ci.yml) uses \u0060pip install -r requirements-dev.txt\u0060 then bare \u0060ruff check .\u0060 and \u0060pytest -q\u0060 \u2014 no \u0060uv run\u0060 anywhere. The README matches the project's actual workflow; the claimed 'uv-based project configuration' is not present in the code."}}, {"stage": "report", "raw_id": "find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "target": "find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "before": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}, "after": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}}, {"stage": "report", "raw_id": "find_removed_behavior:43f62a729df77e4f2a17852c:1", "target": "find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "before": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}, "after": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}}, {"stage": "report", "raw_id": "find_structural:1a47a9cf93ac76ff3b679361:1", "target": "find_structural:1a47a9cf93ac76ff3b679361:1", "before": {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060.", "source": "protopatch", "raw_ids": ["find_structural:1a47a9cf93ac76ff3b679361:1"], "verdict": "confirmed", "note": "pyproject.toml at head has no \u0060dependencies\u0060 key and sets \u0060asyncio_mode = \"auto\"\u0060 (requires pytest-asyncio). However, the project's actual CI (.github/workflows/ci.yml) uses \u0060pip install -r requirements-dev.txt\u0060 + bare \u0060pytest -q\u0060, not \u0060uv run pytest\u0060; the practical impact is limited to anyone who tries the uv workflow."}, "after": {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060.", "source": "protopatch", "raw_ids": ["find_structural:1a47a9cf93ac76ff3b679361:1"], "verdict": "confirmed", "note": "pyproject.toml at head has no \u0060dependencies\u0060 key and sets \u0060asyncio_mode = \"auto\"\u0060 (requires pytest-asyncio). However, the project's actual CI (.github/workflows/ci.yml) uses \u0060pip install -r requirements-dev.txt\u0060 + bare \u0060pytest -q\u0060, not \u0060uv run pytest\u0060; the practical impact is limited to anyone who tries the uv workflow."}}, {"stage": "report", "raw_id": "find_structural:1904e467e1208b0e846f2c3f:1", "before": {"file": "pyproject.toml", "line": 1, "severity": "nit", "category": "conventions", "claim": "README instructs pip-based dev setup (pip install -r requirements-dev.txt, bare ruff/pytest) that conflicts with the uv-based project configuration (uv run ruff, uv run pytest).", "evidence": "README lines 94\u201395 direct developers to \u0060pip install -r requirements-dev.txt\u0060 and then run \u0060ruff\u0060/\u0060pytest\u0060 bare, while the detected project commands all use \u0060uv run\u0060 (e.g. \u0060uv run ruff check .\u0060, \u0060uv run pytest\u0060).", "source": "protopatch", "raw_ids": ["find_structural:1904e467e1208b0e846f2c3f:1"], "verdict": "refuted", "note": "The CI workflow (.github/workflows/ci.yml) uses \u0060pip install -r requirements-dev.txt\u0060 then bare \u0060ruff check .\u0060 and \u0060pytest -q\u0060 \u2014 no \u0060uv run\u0060 anywhere. The README matches the project's actual workflow; the claimed 'uv-based project configuration' is not present in the code."}, "after": {"raw_ids": ["find_structural:1904e467e1208b0e846f2c3f:1"], "disposition": "refuted", "reason": "Verifier confirmed the CI workflow uses pip install -r requirements-dev.txt and bare ruff/pytest, not uv run; the README matches the actual project workflow.", "evidence": "CI workflow (.github/workflows/ci.yml) uses pip install -r requirements-dev.txt then bare ruff check . and pytest -q \u2014 no uv run anywhere."}}, {"stage": "dispatcher", "raw_id": "find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "before": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}, "after": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}}, {"stage": "dispatcher", "raw_id": "find_removed_behavior:43f62a729df77e4f2a17852c:1", "before": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}, "after": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}}, {"stage": "dispatcher", "raw_id": "find_structural:1a47a9cf93ac76ff3b679361:1", "before": {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060.", "source": "protopatch", "raw_ids": ["find_structural:1a47a9cf93ac76ff3b679361:1"], "verdict": "confirmed", "note": "pyproject.toml at head has no \u0060dependencies\u0060 key and sets \u0060asyncio_mode = \"auto\"\u0060 (requires pytest-asyncio). However, the project's actual CI (.github/workflows/ci.yml) uses \u0060pip install -r requirements-dev.txt\u0060 + bare \u0060pytest -q\u0060, not \u0060uv run pytest\u0060; the practical impact is limited to anyone who tries the uv workflow."}, "after": {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060.", "source": "protopatch", "raw_ids": ["find_structural:1a47a9cf93ac76ff3b679361:1"], "verdict": "confirmed", "note": "pyproject.toml at head has no \u0060dependencies\u0060 key and sets \u0060asyncio_mode = \"auto\"\u0060 (requires pytest-asyncio). However, the project's actual CI (.github/workflows/ci.yml) uses \u0060pip install -r requirements-dev.txt\u0060 + bare \u0060pytest -q\u0060, not \u0060uv run pytest\u0060; the practical impact is limited to anyone who tries the uv workflow. \u2014 nearby: in code this PR did not change (outside its changed lines and the functions they sit in) \u2014 reported, not gated (#232)", "nearby": true}}, {"stage": "posted", "raw_id": "find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "before": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}, "after": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}}, {"stage": "posted", "raw_id": "find_removed_behavior:43f62a729df77e4f2a17852c:1", "before": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}, "after": {"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}}, {"stage": "posted", "raw_id": "find_structural:1a47a9cf93ac76ff3b679361:1", "before": {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060.", "source": "protopatch", "raw_ids": ["find_structural:1a47a9cf93ac76ff3b679361:1"], "verdict": "confirmed", "note": "pyproject.toml at head has no \u0060dependencies\u0060 key and sets \u0060asyncio_mode = \"auto\"\u0060 (requires pytest-asyncio). However, the project's actual CI (.github/workflows/ci.yml) uses \u0060pip install -r requirements-dev.txt\u0060 + bare \u0060pytest -q\u0060, not \u0060uv run pytest\u0060; the practical impact is limited to anyone who tries the uv workflow."}, "after": {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060.", "source": "protopatch", "raw_ids": ["find_structural:1a47a9cf93ac76ff3b679361:1"], "verdict": "confirmed", "note": "pyproject.toml at head has no \u0060dependencies\u0060 key and sets \u0060asyncio_mode = \"auto\"\u0060 (requires pytest-asyncio). However, the project's actual CI (.github/workflows/ci.yml) uses \u0060pip install -r requirements-dev.txt\u0060 + bare \u0060pytest -q\u0060, not \u0060uv run pytest\u0060; the practical impact is limited to anyone who tries the uv workflow. \u2014 nearby: in code this PR did not change (outside its changed lines and the functions they sit in) \u2014 reported, not gated (#232)", "nearby": true}}], "complete": true, "losses": [], "report_findings": [{"file": "read_tools.py", "line": 846, "severity": "minor", "category": "correctness", "claim": "github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no longer default to (flagged by both correctness and removed-behavior review).", "evidence": "default, picker = await asyncio.to_thread(lambda: (current_default(default_repo), _repos())) \u2014 the tool's docstring says it reports \"which repo the github_* tools default to\", but resolve_repo now prefers scoped_default() over current_default(default_repo).", "raw_ids": ["find_correctness:9ecb5ed0ed21acf3e8f2e39f:1", "find_removed_behavior:43f62a729df77e4f2a17852c:1"], "verdict": "confirmed", "note": "read_tools.py:846 reads \u0060current_default(default_repo)\u0060 directly; search_pr_repo confirms \u0060scoped_default\u0060 appears nowhere in read_tools.py. resolve_repo (gh_issue.py) now inserts \u0060scoped_default()\u0060 between explicit and \u0060current_default\u0060, so github_status's reported default diverges from what the other tools actually resolve to inside a scoped job."}, {"file": "pyproject.toml", "line": 1, "severity": "minor", "category": "build-release", "claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.", "evidence": "The [project] table (lines 1\u20135) has no \u0060dependencies\u0060 or \u0060optional-dependencies\u0060 key. Line 16 sets \u0060asyncio_mode = \"auto\"\u0060, which requires the \u0060pytest-asyncio\u0060 plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is \u0060uv run pytest\u0060.", "source": "protopatch", "raw_ids": ["find_structural:1a47a9cf93ac76ff3b679361:1"], "verdict": "confirmed", "note": "pyproject.toml at head has no \u0060dependencies\u0060 key and sets \u0060asyncio_mode = \"auto\"\u0060 (requires pytest-asyncio). However, the project's actual CI (.github/workflows/ci.yml) uses \u0060pip install -r requirements-dev.txt\u0060 + bare \u0060pytest -q\u0060, not \u0060uv run pytest\u0060; the practical impact is limited to anyone who tries the uv workflow."}]}There was a problem hiding this comment.
Promoting the PASS verdict for head 8f42d562086b: all checks terminal-green, zero unresolved review threads. (approve-on-green)
Open findings carried by this approval — non-blocking, but they did not go away:
- minor
read_tools.py:846— github_status reports the agent-wide default via current_default(), bypassing the new scoped_default() step in resolve_repo; inside a scoped job it tells the model the tools default to a repo they no - minor
pyproject.toml:1— pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = "auto"), making uv run pytest fail in a clean environment.
Approving a WARN does not resolve its findings (issue #22).
Review finding on #73: github_status reported the configured default even where the tools now resolve an omitted repo to the job's scoped one, telling the model the wrong default. It now appends the scoped default; the operator banner keeps the configured one. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
There was a problem hiding this comment.
QA panel review — PASS
code-review-structural · head e3c1c94c2373 · formal
⚠️ 1 finding(s) carried forward from8f42d562086bstill stand (unchanged files, not re-read this round). Where the brief below says the review is clean, this line supersedes it.
No findings were raised, so there was nothing to verify. The panel's synthesized result was an empty array and the verifier confirmed nothing-to-verify. The one prior minor (read_tools.py:846, scoped-default reporting) remains open from round 1 but is below the blocker/major threshold for a standing disposition. No coverage gaps to flag.
Findings
| Severity | Location | Finding | Verified | |
|---|---|---|---|---|
| 🟡 | minor | pyproject.toml:1 |
pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = "auto"), making uv run pytest fail in a clean environment. | confirmed · nearby, not gating |
findings JSON (machine-readable)
[
{
"file": "pyproject.toml",
"line": 1,
"severity": "minor",
"category": "build-release",
"claim": "pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = \"auto\"), making uv run pytest fail in a clean environment.",
"evidence": "The [project] table (lines 1\u20135) has no `dependencies` or `optional-dependencies` key. Line 16 sets `asyncio_mode = \"auto\"`, which requires the `pytest-asyncio` plugin. The README (line 98) explicitly states the suite is \"Host-free \u2014 the suite needs no protoAgent host,\" and the detected test command is `uv run pytest`.",
"source": "protopatch",
"raw_ids": [
"find_structural:1a47a9cf93ac76ff3b679361:1"
],
"verdict": "confirmed",
"note": "pyproject.toml at head has no `dependencies` key and sets `asyncio_mode = \"auto\"` (requires pytest-asyncio). However, the project's actual CI (.github/workflows/ci.yml) uses `pip install -r requirements-dev.txt` + bare `pytest -q`, not `uv run pytest`; the practical impact is limited to anyone who tries the uv workflow. \u2014 nearby: in code this PR did not change (outside its changed lines and the functions they sit in) \u2014 reported, not gated (#232) \u2014 carried unchanged: this file is byte-identical to the round that raised it (incremental, #313)",
"nearby": true,
"carried": true,
"carried_by": "incremental",
"since": "8f42d562086b6e9a75e74740f0da849dd2b56f11"
}
]1 structural finding(s) are nearby notes, not part of the verdict: they sit in code this PR did not change (outside its changed lines and the functions those lines are in). Worth a look; not a request for this PR.
pyproject.toml:16(minor) — pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = "auto"), making uv run pytest fail in a clean environment.
Runner context: observed. Native parent-conversation exclusion and separate history exclusion are checked per execution. Intended prior findings may remain in the rendered prompt. Instruction-source observations describe the assembled system prompt; mandatory repository guidance remains applicable. External instruction sources and runtime additions remain unknown. Tool observations describe bound tools, inherited fence and enforcement; external policy remains unknown. Construction evidence does not establish semantic review coverage.
Incremental re-review (#313): 2 of 7 file(s) changed since head 8f42d562086b (or depend on a change) and were reviewed. 5 unchanged file(s) were not re-read. 1 finding(s) on them are carried forward unchanged.
Comment @vera review for a full re-review.
Addressed since the last round (1 minor/nit): read_tools.py:846 — no longer raised, and the PR changed the code around them since they were raised.
Raw finding accounting: incomplete
{"supported": true, "binding": {"attempt": "43f820bdccc24bbcb4a2fcb309ac02d4", "repo": "protoLabsAI/github-plugin", "head": "e3c1c94c237316739a8aa4bbd6b6f7d292a0718a", "pr": "73"}, "raw": {}, "lanes": {"find_correctness": {"state": "observed", "count": 0}, "find_removed_behavior": {"state": "observed", "count": 0}, "find_crossfile": {"state": "observed", "count": 0}, "find_conventions": {"state": "observed", "count": 0}, "find_structural": {"state": "observed", "count": 0}}, "gaps": ["verify: missing or malformed findings/accounting array"], "transitions": [], "complete": false, "losses": [], "report_findings": []}There was a problem hiding this comment.
Promoting the PASS verdict for head e3c1c94c2373: all checks terminal-green, zero unresolved review threads. (approve-on-green)
Open findings carried by this approval — non-blocking, but they did not go away:
- minor
pyproject.toml:1— pyproject.toml declares no dependencies but configures pytest-asyncio (asyncio_mode = "auto"), making uv run pytest fail in a clean environment.
Approving a WARN does not resolve its findings (issue #22).
Fixes #72. Consumer: pr-reviewer-plugin#367.
Change
gh_issue.scoped_default_repo(repo): a context manager over aContextVar;gh_issue.scoped_default()reads it.resolve_repo: explicitrepo> scoped default > configured default >GITHUB_DEFAULT_REPO/GH_REPO. Every tool resolves through it, so all read / write / review / release tools honour the scope.Why
Vera runs review rounds across 170 repos with
default_repo: "", so an omitted repo resolves togithub.repos[0]= protoAgent. On mythxengine#966 a finder lane read protoAgent'sCHANGELOG.md/pyproject.tomlthat way (pr-reviewer-plugin#367). The pr-reviewer will wrap each round inscoped_default_repo(round_repo).Verification
Tests: precedence (scope beats configured default, explicit beats scope, nested scopes restore), malformed scopes are ignored, the scope follows
asynciotasks started inside it and not ones outside, andgithub_get_prwithoutrepotargets the scoped repo inside the scope and the configured one after.ruff(0.15.10 pin) +pytest: 579 passed, 13 skipped.🤖 Generated with Claude Code