Add Gitlab artifacts support

skills/smith/SKILL.md

@@ -40,9 +40,11 @@ Full vocabulary and flags live in `references/usage-recipes.md`. The minimum you
 | Discovery | `smith <azdo-remote-name> orgs`, `smith <github-remote-name> orgs`, `smith <gitlab-remote-name> groups`, `smith <azdo-remote-name> repos <project>`, `smith <github-remote-name> repos`, `smith <gitlab-remote-name> repos` |
 | Focused grep | `smith <azdo-remote-name> code grep <project> <repo> "<regex>"`, `smith <github-remote-name> code grep <repo> "<regex>"`, `smith <gitlab-remote-name> code grep <group/project> "<regex>"` |
 | PRs / MRs | `smith <azdo-remote-name> prs search`, `smith <github-remote-name> prs search`, `smith <gitlab-remote-name> prs search`, `smith <github-remote-name> prs list <repo>`, `smith <gitlab-remote-name> prs list <group/project>` |
-| Pipelines | `smith <github-remote-name> pipelines list <repo> <id>`, `smith <gitlab-remote-name> pipelines list <group/project> <id>`, `smith <github-remote-name> pipelines grep <repo> <id> "<regex>"`, `smith <gitlab-remote-name> pipelines grep <group/project> <id> "<regex>"` |
+| Pipelines | `smith <github-remote-name> pipelines list <repo> <id>`, `smith <gitlab-remote-name> pipelines list <group/project> <id>`, `smith <github-remote-name> pipelines grep <repo> <id> "<regex>"`, `smith <gitlab-remote-name> pipelines grep <group/project> <id> "<regex>"`, `smith <gitlab-remote-name> pipelines artifacts list <group/project> <pipeline-id> <job-id>`, `smith <gitlab-remote-name> pipelines artifacts grep <group/project> <pipeline-id> <job-id> "<regex>"` |
 | Stories / Issues | `smith <azdo-remote-name> stories search <project> --query`, `smith <gitlab-remote-name> stories search <group/project> --query`, `smith <youtrack-remote-name> stories search --query` |
 
+All grep commands (code, pipeline logs, artifacts) support: `--path`, `--glob`, `--output-mode` (content/files_with_matches/count), `--context-lines`, `--from-line`/`--to-line`, `--reverse`, `--case-sensitive`. Code grep adds: `--branch`, `--no-clone`. Pipeline grep adds: `--log-id`.
+
 Rules that save retries:
 
 - **GitHub**: repo arg is bare `<repo>`, not `org/repo`. Search output may look like `org/repo:path` but commands still take `<repo>`.
@@ -51,6 +53,7 @@ Rules that save retries:
 - **YouTrack**: no repo arg; only issue IDs (e.g. `RAD-1055`) and `--query`.
 - Global `smith code search` and `smith prs search` target every enabled remote and reject `--project` or `--repo`. Use `smith <remote> ...` to narrow.
 - `pipelines grep ... <id>` expects a pipeline/run/build ID. For a specific job or log, call `pipelines list ...` first to find the parent ID, then `pipelines grep ... <pipeline-id> ".*" --log-id <job-or-log-id>`.
+- `pipelines artifacts ... <pipeline-id> <job-id>` is GitLab-only. Use `artifacts list` to enumerate archive paths and `artifacts grep`
 - `pipelines list ... <id>` prints a compact DAG (`@` pipelines, `#` stages, `*` jobs, inline `<needs` and `>>` downstream). GitLab traverses child pipelines via GraphQL (REST fallback emits header-only rows with a warning). Filter with `--status`, `--grep`, `--skip`/`--take`, `--max-depth` (gitlab only, default 0 = unlimited). Full grammar lives in `references/pipelines-format.md`.
 
 Use `--help` on any command for flags.
@@ -67,6 +70,7 @@ Use `--help` on any command for flags.
 ### Pipeline Analysis
 1. Use `smith pipelines list <repo> <pipeline_id> --status failed` to focus on failed jobs.
 2. Once you know the pipeline log ID, use `smith pipelines grep <repo> <pipeline_id> <log_id> --reverse` to analyze the logs.
+3. For GitLab artifact-backed failures, use `smith <gitlab-remote-name> pipelines artifacts list <group/project> <pipeline_id> <job_id>` and then `... artifacts grep ... "<regex>"`.
 ## Stop Conditions
 
 Stop narrowing and answer when any of these is true:

src/smith/cli/handlers.py

@@ -633,6 +633,47 @@ def handle_ci_grep(client: SmithClient, args: argparse.Namespace) -> int:
     )
 
 
+def handle_ci_artifacts_list(client: SmithClient, args: argparse.Namespace) -> int:
+    data = client.execute_ci_artifacts_list(
+        remote_or_provider=_selected_target(args),
+        project=getattr(args, "project", None),
+        repo=getattr(args, "repo", None),
+        pipeline_id=args.id,
+        job_id=args.job_id,
+    )
+    return _emit_success(
+        args=args,
+        command=args.command_id,
+        data=data,
+        partial=_is_partial_result(data),
+    )
+
+
+def handle_ci_artifacts_grep(client: SmithClient, args: argparse.Namespace) -> int:
+    data = client.execute_ci_artifacts_grep(
+        remote_or_provider=_selected_target(args),
+        project=getattr(args, "project", None),
+        repo=getattr(args, "repo", None),
+        pipeline_id=args.id,
+        job_id=args.job_id,
+        pattern=args.pattern,
+        path=getattr(args, "path", None),
+        glob=getattr(args, "glob", None),
+        output_mode=args.output_mode,
+        case_insensitive=not args.case_sensitive,
+        context_lines=args.context_lines,
+        from_line=args.from_line,
+        to_line=args.to_line,
+        reverse=getattr(args, "reverse", False),
+    )
+    return _emit_success(
+        args=args,
+        command=args.command_id,
+        data=data,
+        partial=_is_partial_result(data),
+    )
+
+
 def handle_work_get(client: SmithClient, args: argparse.Namespace) -> int:
     request_kwargs: dict[str, Any] = {
         "remote_or_provider": _selected_target(args),

src/smith/cli/parser.py

@@ -7,6 +7,8 @@
 from smith.cli.handlers import (
     _csv_list,
     handle_cache_clean,
+    handle_ci_artifacts_grep,
+    handle_ci_artifacts_list,
     handle_ci_grep,
     handle_ci_list,
     handle_code_grep,
@@ -183,6 +185,32 @@ def _add_ci_grep_options(parser: argparse.ArgumentParser) -> None:
     parser.add_argument("--case-sensitive", action="store_true")
 
 
+def _add_artifact_grep_options(parser: argparse.ArgumentParser) -> None:
+    parser.add_argument("--path", help="Artifact path scope within the extracted archive")
+    parser.add_argument("--glob", help="Artifact filename glob filter")
+    parser.add_argument(
+        "pattern",
+        help=(
+            'Regex pattern. Use ".*" to match all. '
+            'Form: smith <remote> pipelines artifacts grep <scope> <pipeline_id> <job_id> "<regex>"'
+        ),
+    )
+    parser.add_argument(
+        "--output-mode",
+        choices=["content", "files_with_matches", "count"],
+        default="content",
+    )
+    parser.add_argument("--context-lines", type=int, default=3)
+    parser.add_argument("--from-line", type=int)
+    parser.add_argument("--to-line", type=int)
+    parser.add_argument(
+        "--reverse",
+        action="store_true",
+        help="Emit matches in reverse order so the most recent hits appear first.",
+    )
+    parser.add_argument("--case-sensitive", action="store_true")
+
+
 def _add_work_search_filters(parser: argparse.ArgumentParser, *, include_area: bool = True) -> None:
     if include_area:
         parser.add_argument("--area")
@@ -549,6 +577,56 @@ def _add_remote_pipelines_group(remote_subparsers: Any, *, remote: RemoteConfig)
     _add_output_format(pipelines_grep)
     _set_handler(pipelines_grep, handle_ci_grep, "pipelines.grep", primary_path="pipelines grep")
 
+    if remote.provider == "gitlab":
+        pipelines_artifacts = _add_parser(
+            pipelines_sub,
+            "artifacts",
+            help_text="List and grep GitLab job artifacts",
+        )
+        pipelines_artifacts_sub = pipelines_artifacts.add_subparsers(
+            dest="pipelines_artifacts_action",
+            required=True,
+        )
+
+        pipelines_artifacts_list = _add_parser(
+            pipelines_artifacts_sub,
+            "list",
+            help_text="List artifact paths for a job",
+        )
+        _add_pipeline_positional_args(
+            pipelines_artifacts_list,
+            remote=remote,
+            id_label=id_label,
+        )
+        pipelines_artifacts_list.add_argument("job_id", type=int, help="Job ID")
+        _add_output_format(pipelines_artifacts_list)
+        _set_handler(
+            pipelines_artifacts_list,
+            handle_ci_artifacts_list,
+            "pipelines.artifacts.list",
+            primary_path="pipelines artifacts list",
+        )
+
+        pipelines_artifacts_grep = _add_parser(
+            pipelines_artifacts_sub,
+            "grep",
+            help_text="Search extracted GitLab job artifacts",
+        )
+        _add_pipeline_positional_args(
+            pipelines_artifacts_grep,
+            remote=remote,
+            id_label=id_label,
+        )
+        pipelines_artifacts_grep.add_argument("job_id", type=int, help="Job ID")
+        _add_artifact_grep_options(pipelines_artifacts_grep)
+        _add_output_format(pipelines_artifacts_grep)
+        _set_handler(
+            pipelines_artifacts_grep,
+            handle_ci_artifacts_grep,
+            "pipelines.artifacts.grep",
+            primary_path="pipelines artifacts grep",
+        )
+
 
 def _add_remote_stories_group(remote_subparsers: Any, *, remote: RemoteConfig) -> None:
     stories = _add_parser(remote_subparsers, "stories", help_text="Get, search, and get mine")

src/smith/client.py

@@ -80,6 +80,15 @@ def _require_single_target(remote_or_provider: str, *, command: str) -> str:
             raise ValueError(f"{command} does not support target 'all'. Use a configured remote name.")
         return target
 
+    def _require_gitlab_target(self, remote_or_provider: str, *, command: str) -> str:
+        target = self._require_single_target(remote_or_provider, command=command)
+        remotes = self._resolve_remotes(target)
+        if not remotes:
+            raise ValueError(f"No enabled remote found for '{target}'")
+        if len(remotes) != 1 or remotes[0].provider != "gitlab":
+            raise ValueError(f"{command} is only supported for GitLab remotes.")
+        return target
+
     def _get_provider_for_remote(self, remote: RemoteConfig) -> BaseProvider:
         if remote.name in self._provider_cache:
             return self._provider_cache[remote.name]
@@ -725,6 +734,74 @@ def execute_ci_grep(
             },
         )
 
+    def execute_ci_artifacts_list(
+        self,
+        *,
+        remote_or_provider: str,
+        project: str | None,
+        repo: str | None,
+        pipeline_id: int,
+        job_id: int,
+    ) -> dict[str, Any]:
+        target = self._require_gitlab_target(
+            remote_or_provider,
+            command="pipelines.artifacts.list",
+        )
+        effective_repo = repo or project
+        return self._fanout(
+            remote_or_provider=target,
+            operations={
+                "gitlab": lambda r: self._gitlab_provider(r).list_job_artifacts(
+                    repo=str(effective_repo),
+                    pipeline_id=pipeline_id,
+                    job_id=job_id,
+                ),
+            },
+        )
+
+    def execute_ci_artifacts_grep(
+        self,
+        *,
+        remote_or_provider: str,
+        project: str | None,
+        repo: str | None,
+        pipeline_id: int,
+        job_id: int,
+        pattern: str | None,
+        path: str | None,
+        glob: str | None,
+        output_mode: Literal["content", "files_with_matches", "count"],
+        case_insensitive: bool,
+        context_lines: int | None,
+        from_line: int | None,
+        to_line: int | None,
+        reverse: bool = False,
+    ) -> dict[str, Any]:
+        target = self._require_gitlab_target(
+            remote_or_provider,
+            command="pipelines.artifacts.grep",
+        )
+        effective_repo = repo or project
+        return self._fanout(
+            remote_or_provider=target,
+            operations={
+                "gitlab": lambda r: self._gitlab_provider(r).grep_job_artifacts(
+                    repo=str(effective_repo),
+                    pipeline_id=pipeline_id,
+                    job_id=job_id,
+                    pattern=pattern,
+                    path=path,
+                    glob=glob,
+                    output_mode=output_mode,
+                    case_insensitive=case_insensitive,
+                    context_lines=context_lines,
+                    from_line=from_line,
+                    to_line=to_line,
+                    reverse=reverse,
+                ),
+            },
+        )
+
     def execute_work_get(
         self,
         *,

src/smith/formatting.py

@@ -239,6 +239,11 @@ def _render_grep(data: Any) -> str:
     return text
 
 
+def _render_artifacts_list(data: Any) -> str:
+    paths = data.get("paths", []) if isinstance(data, dict) else []
+    return "\n".join(str(path) for path in paths if str(path).strip())
+
+
 def _render_cache_clean(data: Any) -> str:
     cleaned = data.get("cleaned", []) if isinstance(data, dict) else []
     missing = data.get("missing", []) if isinstance(data, dict) else []
@@ -758,6 +763,8 @@ def _render_needs(needs: Any, *, name_to_id: dict[str, Any]) -> str:
     "cache.clean": _render_cache_clean,
     "pipelines.list": _render_pipelines_list,
     "pipelines.grep": _render_grep,
+    "pipelines.artifacts.list": _render_artifacts_list,
+    "pipelines.artifacts.grep": _render_grep,
     "prs.search": _render_pr_list,
     "prs.list": _render_pr_list,
     "prs.get": _render_pr_get,
@@ -805,7 +812,7 @@ def _render_remote_grouped(command: str, payload: dict[str, Any]) -> str:
             lines.append(rendered)
 
         warnings = _visible_remote_warnings(command, remote_data, entry.get("warnings") or [])
-        if command not in {"code.grep", "pipelines.grep"}:
+        if command not in {"code.grep", "pipelines.grep", "pipelines.artifacts.grep"}:
             for warning in warnings:
                 lines.append(f"warning: {warning}")
         return "\n".join(lines).rstrip()
@@ -851,7 +858,7 @@ def _render_remote_grouped(command: str, payload: dict[str, Any]) -> str:
                 output_lines.append(rendered)
 
         warnings = _visible_remote_warnings(command, remote_data, entry.get("warnings") or [])
-        if command not in {"code.grep", "pipelines.grep"}:
+        if command not in {"code.grep", "pipelines.grep", "pipelines.artifacts.grep"}:
             for warning in warnings:
                 output_lines.append(f"warning: {warning}")
         output_lines.append("")