Fail on Severity

docs/integrations/ci-cd.mdx

@@ -27,6 +27,12 @@ strix -n --target ./app --scan-mode quick --scope-mode diff --diff-base origin/m
 | 1 | Execution error |
 | 2 | Vulnerabilities found |
 
+To fail only on serious findings, pass `--fail-on` with a minimum severity. Lower findings are still reported but exit `0`:
+
+```bash
+strix -n --target ./app --scan-mode quick --fail-on high
+```
+
 ## GitLab CI
 
 ```yaml .gitlab-ci.yml

docs/integrations/github-actions.mdx

@@ -49,6 +49,8 @@ The workflow fails when vulnerabilities are found:
 | 0 | Pass — No vulnerabilities |
 | 2 | Fail — Vulnerabilities found |
 
+Add `--fail-on high` (or `critical`, `medium`, `low`) to fail only on findings at or above that severity. Lower findings still appear in the report.
+
 ## Scan Modes for CI
 
 | Mode | Duration | Use Case |

docs/usage/cli.mdx

@@ -60,6 +60,14 @@ strix (--target <target> | --target-list <path>) [options]
   Run in headless mode without TUI. Ideal for CI/CD.
 </ParamField>
 
+<ParamField path="--fail-on" type="string">
+  Headless mode only (requires `-n`). Minimum severity that makes the run exit
+  `2`: `critical`, `high`, `medium`, `low`, or `info`. Findings below the
+  threshold are still written to every report artifact. A finding whose
+  severity Strix does not recognize always counts, so the gate never passes on a
+  value it cannot rank. Omit the flag to exit `2` on any finding.
+</ParamField>
+
 <ParamField path="--config" type="string">
   Path to a custom config file (JSON) to use instead of `~/.strix/cli-config.json`.
 </ParamField>
@@ -132,6 +140,9 @@ strix --target api.example.com --instruction "Focus on IDOR and auth bypass"
 # CI/CD mode
 strix -n --target ./ --scan-mode quick
 
+# CI/CD mode, failing only on high or critical findings
+strix -n --target ./ --scan-mode quick --fail-on high
+
 # Cap cost and per-agent turns
 strix --target https://example.com --max-budget 25 --max-turns 300
 
@@ -161,4 +172,4 @@ strix --target https://app.com --workspace-file ./openapi.yaml:specs/openapi.yam
 |------|---------|
 | 0 | Scan completed successfully (interactive mode always exits `0`; in headless mode, `0` means no vulnerabilities were found) |
 | 1 | A fatal error occurred before or during the scan (e.g. missing environment variables, Docker unavailable, invalid config file, diff-scope resolution failure, or an unhandled error) |
-| 2 | Vulnerabilities found (headless mode only) |
+| 2 | Vulnerabilities found (headless mode only; with `--fail-on`, only findings at or above that severity) |

skills/ci-security-scanning-with-strix/SKILL.md

@@ -67,7 +67,7 @@ Then tell the user to add two repository secrets: `STRIX_LLM` (model id, for exa
 
 Notes:
 - In CI/headless runs Strix automatically scopes to the PR's changed files (`--scope-mode auto`). If diff resolution fails, keep `fetch-depth: 0` or set `--diff-base` to the PR's actual base branch — use `origin/${{ github.base_ref }}` in GitHub Actions rather than a hard-coded `origin/main`, since repos use different default branches.
-- Exit codes: `0` pass, `2` vulnerabilities found (fails the job), `1` setup error.
+- Exit codes: `0` pass, `2` vulnerabilities found (fails the job), `1` setup error. Add `--fail-on high` to fail only on high/critical findings; lower ones are still reported.
 - The runner needs Docker (default GitHub-hosted Ubuntu runners have it).
 - **Size the budget so the scan completes — do not let it fail open.** A `0` exit means "no validated vulnerabilities in what was analyzed"; if `--max-budget` is hit before the diff is fully covered, the scan wraps up early and can still exit `0`. The "Fail unless the scan completed" step above narrows the gap: `strix_runs/<run>/run.json` is `"stopped"` when the scan was cut off at the hard budget limit without a final report. It is not a complete guard — the agents get graduated wrap-up warnings before that limit, and a run that wraps up on a warning still calls `finish_scan` and records `"completed"` with partial coverage. So keep that step in any pipeline that gates merges **and** give the scan real headroom (compare `run.json`'s `llm_usage.cost` against `--max-budget`; if it ran right up to the cap, raise it). For a `quick` diff-scoped PR scan `--max-budget 10` is usually ample, raise it for large diffs.
 

skills/penetration-testing-with-strix/SKILL.md

@@ -94,6 +94,7 @@ Key flags:
 | `--workspace-file PATH[:DEST]` | Copy a file from this machine into `/workspace` before the scan, for a wordlist, a spec, or notes. Repeatable. |
 | `--max-budget USD` | Hard LLM spend cap; scan wraps up cleanly at the limit. |
 | `--max-turns N` | Per-agent turn cap (default 500). |
+| `--fail-on SEVERITY` | Headless only: exit `2` only for findings at or above `critical`/`high`/`medium`/`low`/`info`. Default: any finding. |
 | `--resume RUN_NAME` | Resume a prior run from `strix_runs/`, with its agent history and targets. Cannot be combined with `-t`. |
 | `--scope-mode` | For code targets: `auto` (diff-scope in CI/headless), `diff` (force changed files only), `full` (whole tree). |
 | `--diff-base REF` | Branch or commit that `diff` scope compares against. Defaults to the repo's default branch. |
@@ -104,7 +105,7 @@ Scans take minutes (`quick`) to hours (`deep`). Run them in the background and p
 
 - `0` — finished with no validated vulnerabilities **in what was analyzed**
 - `1` — fatal error (missing env vars, Docker down, bad config)
-- `2` — vulnerabilities found
+- `2` — vulnerabilities found (with `--fail-on`, only at or above that severity; lower findings still land in the artifacts)
 
 A `0` is not proof of full coverage: if `--max-budget`/`--max-turns` is reached before the scan completes, it wraps up early and still exits `0`. When you need assurance the scan finished, give it enough budget and check `strix_runs/<run>/run.json`: a hard budget stop leaves `status: "stopped"`, but an agent that wrapped up early on a budget *warning* still calls `finish_scan` and records `"completed"` — so also sanity-check the run's cost against `--max-budget` and the report's stated coverage before treating a clean result as full coverage.
 

strix/interface/cli_args.py

@@ -20,6 +20,10 @@ from strix.interface.utils import (
 )
 
 
+# Severities ``--fail-on`` accepts, most severe first.
+FAIL_ON_SEVERITIES = ("critical", "high", "medium", "low", "info")
+
+
 def get_version() -> str:
     try:
         from importlib.metadata import version
@@ -186,6 +190,21 @@ Strix Cloud:
         ),
     )
 
+    parser.add_argument(
+        "--fail-on",
+        dest="fail_on",
+        type=str.lower,
+        choices=FAIL_ON_SEVERITIES,
+        default=None,
+        metavar="SEVERITY",
+        help=(
+            "Headless mode only: exit 2 only when a finding is at or above this severity "
+            "(critical, high, medium, low, info). Lower findings are still written to every "
+            "report artifact. A finding with an unrecognized severity always counts. "
+            "Default: any finding exits 2."
+        ),
+    )
+
     parser.add_argument(
         "-m",
         "--scan-mode",
@@ -318,6 +337,9 @@ Strix Cloud:
     if args.update:
         sys.exit(0 if self_update() else 1)
 
+    if args.fail_on and not args.non_interactive:
+        parser.error("--fail-on only applies to headless runs; add -n/--non-interactive.")
+
     if args.instruction and args.instruction_file:
         parser.error(
             "Cannot specify both --instruction and --instruction-file. Use one or the other."

strix/interface/main.py

@@ -8,6 +8,7 @@ import asyncio
 import contextlib
 import sys
 from pathlib import Path
+from typing import Any
 
 from rich.console import Console
 from rich.panel import Panel
@@ -15,7 +16,7 @@ from rich.text import Text
 
 from strix.config import codex, load_settings, persist_current
 from strix.core.paths import run_dir_for
-from strix.interface.cli_args import parse_arguments
+from strix.interface.cli_args import FAIL_ON_SEVERITIES, parse_arguments
 from strix.interface.environment import (
     check_docker_installed,
     pull_docker_image,
@@ -315,6 +316,30 @@ def display_completion_message(args: argparse.Namespace, results_path: Path) ->
         notify_update(console)
 
 
+def findings_fail_build(reports: list[dict[str, Any]], fail_on: str | None) -> bool:
+    """Whether headless findings should exit 2 under the ``--fail-on`` threshold.
+
+    With no threshold any finding fails. Otherwise a finding fails when its
+    severity is at or above the threshold. A severity outside the known scale
+    fails too, so a gate never passes on a value it cannot rank. ``none`` is a
+    known level below ``info`` and only fails without a threshold.
+    """
+    if not reports:
+        return False
+    if fail_on is None:
+        return True
+    threshold = FAIL_ON_SEVERITIES.index(fail_on)
+    for report in reports:
+        severity = str(report.get("severity") or "").strip().lower()
+        if severity == "none":
+            continue
+        if severity not in FAIL_ON_SEVERITIES:
+            return True
+        if FAIL_ON_SEVERITIES.index(severity) <= threshold:
+            return True
+    return False
+
+
 def _print_error_panel(title: str, message: str) -> None:
     console = Console()
     error_text = Text()
@@ -501,7 +526,7 @@ def main() -> None:
 
     if args.non_interactive:
         report_state = get_global_report_state()
-        if report_state and report_state.vulnerability_reports:
+        if report_state and findings_fail_build(report_state.vulnerability_reports, args.fail_on):
             sys.exit(2)
 
 

tests/test_cli_fail_on.py

@@ -0,0 +1,112 @@
+"""Tests for the --fail-on headless severity gate."""
+
+from __future__ import annotations
+
+import importlib
+import sys
+from types import SimpleNamespace
+from typing import Any
+
+import pytest
+
+
+cli_main: Any = importlib.import_module("strix.interface.main")
+
+
+def _stub_settings(monkeypatch: pytest.MonkeyPatch) -> None:
+    monkeypatch.setattr(
+        cli_main,
+        "load_settings",
+        lambda: SimpleNamespace(runtime=SimpleNamespace(max_local_copy_mb=1024)),
+    )
+
+
+def _reports(*severities: str | None) -> list[dict[str, Any]]:
+    return [{"id": f"vuln-{i:04d}", "severity": sev} for i, sev in enumerate(severities, 1)]
+
+
+def test_no_findings_never_fail() -> None:
+    assert not cli_main.findings_fail_build([], None)
+    assert not cli_main.findings_fail_build([], "info")
+
+
+@pytest.mark.parametrize("severity", ["critical", "high", "medium", "low", "info", "none"])
+def test_without_threshold_any_finding_fails(severity: str) -> None:
+    assert cli_main.findings_fail_build(_reports(severity), None)
+
+
+@pytest.mark.parametrize(
+    ("fail_on", "severity", "expected"),
+    [
+        ("high", "critical", True),
+        ("high", "high", True),
+        ("high", "medium", False),
+        ("high", "low", False),
+        ("high", "info", False),
+        ("critical", "high", False),
+        ("critical", "critical", True),
+        ("info", "info", True),
+        ("info", "low", True),
+        ("medium", "HIGH", True),
+        ("medium", " Low ", False),
+    ],
+)
+def test_threshold_compares_severity(fail_on: str, severity: str, expected: bool) -> None:
+    assert cli_main.findings_fail_build(_reports(severity), fail_on) is expected
+
+
+def test_one_finding_at_threshold_fails_a_mixed_run() -> None:
+    assert cli_main.findings_fail_build(_reports("info", "low", "high"), "high")
+
+
+@pytest.mark.parametrize("severity", ["severe", "", None])
+def test_unrecognized_severity_fails_closed(severity: str | None) -> None:
+    assert cli_main.findings_fail_build(_reports(severity), "critical")
+
+
+def test_none_severity_passes_any_threshold() -> None:
+    assert not cli_main.findings_fail_build(_reports("none"), "info")
+
+
+def test_parse_fail_on_is_case_insensitive(monkeypatch: pytest.MonkeyPatch) -> None:
+    _stub_settings(monkeypatch)
+    monkeypatch.setattr(
+        sys, "argv", ["strix", "-t", "https://test.com/", "-n", "--fail-on", "HIGH"]
+    )
+
+    args = cli_main.parse_arguments()
+
+    assert args.fail_on == "high"
+
+
+def test_parse_fail_on_defaults_to_none(monkeypatch: pytest.MonkeyPatch) -> None:
+    _stub_settings(monkeypatch)
+    monkeypatch.setattr(sys, "argv", ["strix", "-t", "https://test.com/", "-n"])
+
+    assert cli_main.parse_arguments().fail_on is None
+
+
+def test_parse_fail_on_rejects_unknown_severity(
+    monkeypatch: pytest.MonkeyPatch, capsys: pytest.CaptureFixture[str]
+) -> None:
+    _stub_settings(monkeypatch)
+    monkeypatch.setattr(
+        sys, "argv", ["strix", "-t", "https://test.com/", "-n", "--fail-on", "severe"]
+    )
+
+    with pytest.raises(SystemExit):
+        cli_main.parse_arguments()
+
+    assert "--fail-on" in capsys.readouterr().err
+
+
+def test_parse_fail_on_requires_non_interactive(
+    monkeypatch: pytest.MonkeyPatch, capsys: pytest.CaptureFixture[str]
+) -> None:
+    _stub_settings(monkeypatch)
+    monkeypatch.setattr(sys, "argv", ["strix", "-t", "https://test.com/", "--fail-on", "high"])
+
+    with pytest.raises(SystemExit):
+        cli_main.parse_arguments()
+
+    assert "--fail-on only applies to headless runs" in capsys.readouterr().err