Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@src/snakemake/workflow.py`:
- Around line 293-297: The function info_conda_env directly indexes os.environ
causing a KeyError when no conda env is active; change info_conda_env to use
os.environ.get("CONDA_DEFAULT_ENV") and os.environ.get("CONDA_PREFIX") (or check
membership) to safely build the string, treat missing/empty values as absent and
return "n/a" when both are missing or result in " ()", and ensure info_header
(where info_conda_env is used) receives a stable string instead of raising.
- Line 314: The header currently references the global `workflow.configfiles`;
change it to use the instance attribute `self.configfiles` instead (update the
code that emits "Config file(s): {workflow.configfiles}" to use
`self.configfiles`) so the message reflects the current Workflow instance rather
than a mutable global; locate the string construction in the Workflow class
(method that generates headers in src/snakemake/workflow.py) and replace the
global reference with the instance property `self.configfiles`.
- Around line 1311-1315: The logger.info call for LogEvent.WORKFLOW_STARTED uses
a misspelled key and the wrong Snakefile value; change the metadata key from
worklow_id to workflow_id and supply the workflow's main Snakefile path instead
of self.snakefile (e.g. use the attribute that holds the workflow’s main
Snakefile such as self.workflow.snakefile or self.main_snakefile) in the extra
payload for the logger.info(self.info_header, extra=...) that emits
LogEvent.WORKFLOW_STARTED.

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@src/snakemake/workflow.py`:
- Line 280: The code uses os.getlogin() to populate "user", which can raise
OSError in headless environments; replace it with a robust fallback such as
using getpass.getuser() (or wrap os.getlogin() in a try/except and fall back to
getpass.getuser() or os.environ.get("USER","unknown")) so the "user" field is
reliably populated; update the site that constructs the dict (the line with
"user": os.getlogin()) to use the new fallback approach.

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@src/snakemake/workflow.py`:
- Around line 283-285: Replace the shell-based call populating the
"conda_version" metadata (currently using subprocess.getoutput("conda
--version").removeprefix("conda ")) with a non-shell probe: first check for
conda presence using shutil.which("conda"), and if found call
subprocess.run(["conda", "--version"], capture_output=True, text=True,
check=False) to read stdout and strip the "conda " prefix; if not found or the
call fails, set the value to an empty string (or None) to preserve graceful
fallback. Update the code that assigns the "conda_version" key accordingly
(refer to the "conda_version" dict key and the surrounding metadata assembly) to
match the project's existing pattern used in shell.py / singularity.py / cwl.py.

---

Duplicate comments:
In `@src/snakemake/workflow.py`:
- Around line 1294-1296: The event is populating the snakefile field with
self.snakefile (the call-site file) instead of the user's main Snakefile; update
the event construction to use self.main_snakefile for the snakefile field (i.e.,
replace snakefile=self.snakefile with snakefile=self.main_snakefile) so the
emitted event (alongside snakefile_main=self.main_snakefile and
workflow_id=uuid.uuid4()) correctly reports the user workflow file.
- Around line 287-288: The code currently indexes os.environ for
CONDA_DEFAULT_ENV and CONDA_PREFIX (setting "conda_env" and "conda_prefix")
which raises KeyError outside a conda environment; change those lookups to safe
retrieval (e.g., os.environ.get or os.getenv) and/or default to None/empty
string so startup won't crash; update the places that set "conda_env" and
"conda_prefix" in the workflow creation to use the safe getter (references:
"conda_env", "conda_prefix", and os.environ usage in the workflow code).

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@src/snakemake/logging.py`:
- Around line 161-183: format_workflow_started currently calls "".format(**msg)
which raises KeyError if any expected key is missing; to fix, merge msg with a
dictionary of safe default values (e.g., empty strings or placeholders) before
formatting so all keys used in the template are present, then call the format on
that merged mapping; update format_workflow_started to copy/merge the incoming
msg and setdefaults for keys like datetime, workflow_id, platform, host, user,
snakemake_version, python_version, conda_version, conda_env, conda_prefix, cmd,
snakefile_main, basedir, rundir, cwd, configfiles so logging remains resilient
to schema drift.

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@src/snakemake/workflow.py`:
- Around line 285-287: The subprocess call uses a hardcoded "conda" which can be
ambiguous and trigger security warnings; capture the resolved path from
shutil.which and pass that to subprocess.run instead. Change the code that
checks shutil.which("conda") to assign its result (e.g., via a walrus operator
into conda_path or by setting conda_path before the subprocess call) and then
call subprocess.run with [conda_path, "--version"] (reference symbols:
shutil.which, subprocess.run, conda_path) to avoid hardcoded paths and duplicate
lookups.

---

Duplicate comments:
In `@src/snakemake/workflow.py`:
- Around line 298-302: The startup logging accesses
os.environ["CONDA_DEFAULT_ENV"] and os.environ["CONDA_PREFIX"] guarded only by
shutil.which("conda"), which can still raise KeyError if no conda env is
activated; change those lookups in the block that builds the dict (the entries
labeled "CONDA_DEFAULT_ENV"/"conda_prefix") to use
os.environ.get("CONDA_DEFAULT_ENV", "n/a") and os.environ.get("CONDA_PREFIX",
"n/a") (or os.getenv) so they safely fall back to "n/a" even when the keys are
missing.