Repository navigation
Conversation
- Add Makefile with uv sync --dev for dependency management - Add pre-commit hooks that run uv format on every commit - Add pre-commit dependency to dev group
- Add uv version check requiring >= 0.8.13 - Replace multiple targets with single 'build' target - Use native 'uv format' command instead of ruff directly - Reduce from 8 targets to 5 for simplicity
malhotra5
pushed a commit
that referenced
this pull request
Sep 6, 2025
… requirements - Test requirement #1: Multiple pause method calls successively only create one PauseEvent - Test requirement #2: Calling conversation.pause() while conversation.run() is running in separate thread will pause the agent - Test requirement #3: Calling conversation.run() on an already paused agent will resume it The tests now properly validate the actual pause/resume behavior where run() resets the pause flag and resumes execution. Co-authored-by: openhands <openhands@all-hands.dev>
8 tasks done
3 tasks done
6 tasks done
2 of 5 tasks
9 tasks done
1 of 2 tasks
This was referenced Jan 29, 2026
xingyaoww
pushed a commit
that referenced
this pull request
Feb 9, 2026
This commit addresses the critical issues raised in the code review: ## Critical Issue #1: Data Structure Design Flaw - Refactored AgentState to AgentStateRegistry using dict[str, Any] pattern - Provides loosely-coupled storage for agent-specific state - Agents store state using string keys instead of typed fields - Maintains backward compatibility with AgentState alias ## Critical Issue #2: Hidden State Mutation - Fixed _check_iterative_refinement to only increment iteration counter when refinement will actually continue - State mutation now happens AFTER validation checks pass - Function name now accurately reflects its behavior ## Critical Issue #3: Missing Tests - Added comprehensive tests for AgentStateRegistry (14 tests) - Added tests for IterativeRefinementConfig validation (4 tests) - Added tests for _check_iterative_refinement logic (12 tests) - Tests cover edge cases: max iterations, threshold boundaries, etc. ## Suggestion #5: Unnecessary Forward Reference - Removed quotes from LLM type hint in example ## Suggestion #7: Unused Method - Removed reset_iterative_refinement() as part of registry refactor ## Note on Issue #4 (FinishTool Special Case) The suggestion to move iterative refinement to the conversation layer is a valid architectural improvement but requires more significant changes. This is deferred to a future PR to keep this change focused. Co-authored-by: openhands <openhands@all-hands.dev>
This was referenced Feb 9, 2026
5 tasks done
5 tasks done
1 task done
2 of 3 tasks
This was referenced Jun 16, 2026
This was referenced Jul 6, 2026
Merged
jpshackelford
pushed a commit
that referenced
this pull request
Jul 8, 2026
Fixes three issues when loading root SKILL.md plugins: Issue #1 (🔴): Malformed root .mcp.json silently dropped the entire skill - Plugin-level loader tolerates malformed .mcp.json (logs warning) - Skill-level loader raised SkillValidationError, caught by broad except - Result: skill silently dropped (0 skills loaded) Issue #2 (🟠): Root .mcp.json parsed twice with divergent expansion - Plugin level: expand_defaults=False (preserves ${VAR:-default}) - Skill level: expand_defaults=True (eagerly expands defaults) - Result: same file processed with different semantics Issue #3 (🟡): Redundant resource discovery - discover_skill_resources() called by both Skill.load() and plugin loader - Minor performance issue, but unnecessary duplication Root cause: For root SKILL.md plugins, plugin_dir == skill_root, so .mcp.json at plugin root is the same file both loaders attempt to process. Solution: - Add skip_mcp parameter to Skill.load() (defaults to False) - Pass skip_mcp=True when loading root skills from plugin loader - Remove redundant discover_skill_resources() calls from plugin loader - Add comprehensive regression tests (4 new tests) Test coverage: - test_malformed_root_mcp_json_does_not_drop_skill - test_root_mcp_json_not_double_loaded - test_root_skill_resources_not_duplicated - test_nested_skill_with_own_mcp_json_still_loads All 266 tests pass (46 plugin + 220 skill tests) Co-authored-by: openhands <openhands@all-hands.dev>
1 of 5 tasks
3 of 5 tasks
1 task done
1 of 5 tasks
This was referenced Jul 30, 2026
1 of 5 tasks
1 of 5 tasks
1 of 6 tasks
Closed
1 of 5 tasks
Open
9 tasks
2 of 5 tasks
This was referenced Oct 1, 2026
This was referenced Oct 11, 2026
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.
No description provided.