Fix linter errors and enable static analysis in CI

[Copilot]

The linter was disabled in CI due to ~800 errors. This PR fixes all linting issues and re-enables hatch fmt --check in the workflow.

Configuration

• Added [tool.ruff] config to pyproject.toml with target Python 3.11
• Suppressed overly strict rules (relative imports, print in CLI tools, exception message formatting, etc.)
• Added per-file ignores for release_tools/, tests/, and forked jsonrpyc/

Bug Fixes

• Added missing logging import in codeql/client.py (F821)
• Removed duplicate RunHooks import in agent.py (F811)
• Removed unused io import in jsonrpyc/__init__.py (F401)
• Fixed unused loop variables with dict.values()/dict.keys() where applicable

Formatting

• Applied ruff formatter to 24 files for consistent style

CI

- name: Run static analysis
  run: hatch fmt --check

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Fix linter errors</issue_title>
<issue_description>I switched off the linter because it produces 100s of error messages:

seclab-taskflow-agent/.github/workflows/ci.yml

Line 38 in d4d4acd

# hatch fmt --check

We should fix those errors (or find a way to suppress them if they're uninteresting), so that we can switch the check on.</issue_description>

<agent_instructions>Run hatch fmt --check and fix all the errors that it finds.</agent_instructions>

Comments on the Issue (you are @copilot in this section)

• Fixes Fix linter errors #83

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

To fix this issue, ensure that all return points of parse_prompt_args return tuples of the same length. The best fix, maintaining existing semantics, is to pad the returned tuple at line 93 (return None, None, None, None, help_msg) by adding another None before help_msg, making it a 6-tuple. This will match the number and order of values returned in other exit points at line 104 and 108. Evaluate whether this impacts any downstream logic (callers expecting a short tuple), but as per instruction, remain within the scope of the shown code.

No additional methods or imports are needed. Only a change to line 93 is required in src/seclab_taskflow_agent/main.py.

To fix the problem, remove the line assigning to _name, since the variable is unused and its assignment does not have side effects. This means deleting line 484:

484:             _name = task_body.get("name", "taskflow")  # placeholder, not used yet

No additional imports, method changes, or variable definitions are required. The fix is a single-line deletion inside the provided code region of src/seclab_taskflow_agent/__main__.py.

The best way to fix this is to replace the property definition for closed in the OutputStream protocol from:

@property
def closed(self) -> bool: ...

to use pass as the body:

@property
def closed(self) -> bool:
    pass

Alternatively, if you wish to be more explicit, you might raise NotImplementedError, but for Protocol definitions with no provided implementation, pass is preferred for clearer intent and to avoid unnecessary runtime errors. Only lines 51-52 in the file need to be changed. There are no new imports or other code changes required.

The best way to fix this issue is to replace the standalone ellipsis (...) as the method body for flush(self) in the OutputStream protocol with a single pass statement. This makes it clear that the method does nothing at runtime and is intended only to specify the required interface. Edit the region in file src/seclab_taskflow_agent/mcp_servers/codeql/jsonrpyc/__init__.py at the definition of the flush method of the OutputStream protocol (on or around line 56) and replace the line containing ... with pass. No additional imports, method definitions, or other changes are required.