feat: enforce runtime agent budgets - #37
Conversation
|
Warning Review limit reached
Next review available in: 17 minutes You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. How can I continue?After more reviews become available, a review can be triggered using the To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews. How do review limits work?CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability. For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window. Please refer docs for additional details. Review details⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (5)
📝 WalkthroughWalkthroughThe PR adds a standard-library runtime guard for dispatch validation, telemetry, budgets, repairs, audits, persistence, and CLI control. It expands Sol/Luna/Terra contracts and documentation, strengthens repository validation, and adds comprehensive tests. ChangesRuntime-guarded router
Estimated code review effort: 4 (Complex) | ~60 minutes Sequence Diagram(s)sequenceDiagram
participant SolPlanner
participant RuntimeGuard
participant LunaWorker
participant TerraAuditor
SolPlanner->>RuntimeGuard: initialize dispatch with identities and budgets
RuntimeGuard->>LunaWorker: permit bounded execution
LunaWorker->>RuntimeGuard: record event telemetry
RuntimeGuard->>TerraAuditor: route exhaustion or register revision audit
TerraAuditor->>RuntimeGuard: report audit completion and claims
Possibly related PRs
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches 💡 1📝 Generate docstrings 💡
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (3)
scripts/validate_repo.py (1)
468-479: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low valueConsider AST-based checks instead of raw source substrings.
Markers such as
'sub.add_parser("start"'depend on the local variable namesub, on double quotes, and on single-line formatting. A formatter run or a rename inruntime_guard.pybreaks repository validation without any behavior change. Detect the subcommands and the named constants through the parsedtreeinstead, and keep substring checks only for message literals.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@scripts/validate_repo.py` around lines 468 - 479, Replace the raw-source checks in the runtime-gate validation loop with AST-based detection using the parsed tree: identify registered subcommands independent of the parser variable, quote style, and formatting, and detect the named constants by their AST definitions or references. Retain substring checks only for required message literals, while preserving the existing missing-gate errors.tests/test_runtime_guard.py (1)
74-78: 🩺 Stability & Availability | 🔵 Trivial | ⚡ Quick winAdd a
timeoutto eachsubprocess.runcall.Each call pipes JSON to the guard CLI and waits for exit. If the CLI blocks on stdin or loops, the test suite hangs with no output instead of failing. A timeout keeps the failure bounded and diagnosable.
♻️ Example change
result = subprocess.run( [sys.executable, "-B", str(GUARD_PATH), "schema"], - text=True, capture_output=True, check=False, + text=True, capture_output=True, check=False, timeout=30, )Note on the static analysis hints for these lines: the
S603andsubprocess-from-requestfindings are false positives. The argument vector is a fixed list built fromsys.executableand a repository path, and no shell is used.Also applies to: 98-106, 111-121, 122-128
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@tests/test_runtime_guard.py` around lines 74 - 78, Add an explicit finite timeout argument to every subprocess.run call in the runtime guard tests, including the calls covering the schema, stdin JSON, and related CLI cases. Keep the existing fixed argument vectors and result assertions unchanged so blocked or looping guard processes fail promptly.Source: Linters/SAST tools
tests/test_validate_repo.py (1)
229-241: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winMake the negative case independent of the marker occurrence count, and widen the branch coverage.
source.replace("duplicate_audit_revision", "removed", 1)removes only the first occurrence.validate_runtime_guard()usesif term not in source, so the test only passes while that literal appears exactly once inruntime_guard.py. If a second occurrence is added later, this test fails for a reason unrelated to the contract. Drop the count argument.The test also leaves three branches of
validate_runtime_guard()uncovered: the unreadable-file path, theSyntaxErrorpath, and the import allowlist. Each is cheap to exercise with a temporary file.Minor naming point: the test name states "executable", but the test does not run the guard.
♻️ Proposed changes
validate_repo.ERRORS.clear() self.write( ".agents/skills/lean-dev-router/scripts/runtime_guard.py", - source.replace("duplicate_audit_revision", "removed", 1), + source.replace("duplicate_audit_revision", "removed"), ) validate_repo.validate_runtime_guard() self.assertTrue(any("duplicate_audit_revision" in error for error in validate_repo.ERRORS)) + + def test_runtime_guard_rejects_bad_syntax_and_foreign_imports(self) -> None: + relative = ".agents/skills/lean-dev-router/scripts/runtime_guard.py" + self.write(relative, "def broken(:\n") + validate_repo.validate_runtime_guard() + self.assertTrue(any("invalid Python" in error for error in validate_repo.ERRORS)) + + validate_repo.ERRORS.clear() + self.write(relative, "import requests\n") + validate_repo.validate_runtime_guard() + self.assertTrue(any("requests" in error for error in validate_repo.ERRORS))🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@tests/test_validate_repo.py` around lines 229 - 241, Update test_runtime_guard_is_executable_and_required to replace every duplicate_audit_revision occurrence in the negative fixture, avoiding dependence on marker count. Extend coverage for validate_runtime_guard with temporary-file cases for unreadable files, SyntaxError input, and import allowlist validation, and rename the test to reflect that it validates required content rather than execution.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.agents/skills/lean-dev-router/scripts/runtime_guard.py:
- Around line 436-454: Add an abandon_audit terminal transition to the audit
lifecycle, validating PLAN_ID, DISPATCH_ID, REVISION, AUDITOR_INSTANCE_ID, and
TERMINATION_REASON, locating the running job, recording status "abandoned" and
its termination reason, and returning the parent:sol decision. Wire the
audit_job action routing to invoke it, release the running-job block in
begin_audit, and leave last_completed_audits unchanged so the next audit
requires full mode.
---
Nitpick comments:
In `@scripts/validate_repo.py`:
- Around line 468-479: Replace the raw-source checks in the runtime-gate
validation loop with AST-based detection using the parsed tree: identify
registered subcommands independent of the parser variable, quote style, and
formatting, and detect the named constants by their AST definitions or
references. Retain substring checks only for required message literals, while
preserving the existing missing-gate errors.
In `@tests/test_runtime_guard.py`:
- Around line 74-78: Add an explicit finite timeout argument to every
subprocess.run call in the runtime guard tests, including the calls covering the
schema, stdin JSON, and related CLI cases. Keep the existing fixed argument
vectors and result assertions unchanged so blocked or looping guard processes
fail promptly.
In `@tests/test_validate_repo.py`:
- Around line 229-241: Update test_runtime_guard_is_executable_and_required to
replace every duplicate_audit_revision occurrence in the negative fixture,
avoiding dependence on marker count. Extend coverage for validate_runtime_guard
with temporary-file cases for unreadable files, SyntaxError input, and import
allowlist validation, and rename the test to reflect that it validates required
content rather than execution.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro Plus
Run ID: c0289756-c485-4697-98da-fcaa7adc3a6f
📒 Files selected for processing (11)
.agents/skills/lean-dev-router/SKILL.md.agents/skills/lean-dev-router/scripts/runtime_guard.pyREADME.mdagents/luna-worker.tomlagents/sol-planner.tomlagents/terra-auditor.tomldocs/zh-CN/README.mdlean-dev-router-self-test-guide.mdscripts/validate_repo.pytests/test_runtime_guard.pytests/test_validate_repo.py
1 participant
概要
独立实现 issue #36：为 Lean Dev Router 增加运行时预算与硬熔断机制。
本 PR 直接基于
main，不依赖 PR #35，也不包含terra_planner快路径或第四角色改动。主要改动
runtime_guard.py，在启动子代理前校验完整 DISPATCH。spinning检测、升级锁存、角色身份租约和 parent 禁止写入门禁。验证
python -B scripts/validate_repo.py：通过origin/mainCloses #36
Summary by CodeRabbit
New Features
Documentation
Validation & Tests