feat(secrets): webhook URL and git tokens sealed; run output masks its secrets (#5037)

#5037 "Secrets at rest and in run output: webhook URL and git token
encrypted, askpass via env, secret values scrubbed from logs".

- webhook.url joins SECRET_KEYS; existing values are sealed at startup by
  migrate_plaintext_secrets. Settings -> Notifications no longer echoes the
  URL: it names the saved host, keeps it when left blank, and has a Remove
  box.
- Git source http_token is sealed inside ansible.sources on save and at
  startup, opened in get_sources, and counted by the undecryptable-secrets
  scan as ansible.sources[<name>].http_token.
- The askpass script holds no token: git runs it with STEWARD_GIT_TOKEN in
  its environment and printf prints it as data.
- The executor masks every secret the run is handed (become/vault
  passwords, bootstrap passwords, sensitive run vars, and the per-host agent
  tokens the host_agent plugin now passes as redact) in each line before it
  is broadcast, logged or stored.
- Bundled "Write agent config" tasks are no_log; a test fails if any bundled
  task templating steward_token lacks it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

steward/ansible/executor.py

@@ -293,6 +293,45 @@ def cancel_run(run_id: str) -> bool:
 
 # ── Run execution ─────────────────────────────────────────────────────────────
 
+# ── Secret scrubbing ──────────────────────────────────────────────────────────
+# Ansible echoes values back: a failed task prints its args, -v prints vars,
+# --diff prints templated files. Each line is masked before it is broadcast,
+# written to the log artifact or stored on the run, so a secret the run was
+# handed never leaves it.
+MASK = "***"
+_MIN_SECRET_LEN = 4  # masking "on" or "1" would mangle the output and hide nothing
+
+
+def _strings(value) -> list[str]:
+    if isinstance(value, str):
+        return [value]
+    if isinstance(value, dict):
+        return [s for v in value.values() for s in _strings(v)]
+    if isinstance(value, (list, tuple, set)):
+        return [s for v in value for s in _strings(v)]
+    return []
+
+
+def run_secrets(creds: dict | None, connection: dict | None,
+                secret_vars: dict | None, redact: list[str] | None) -> list[str]:
+    """Every secret value this run is handed: the global become/vault
+    passwords, the bootstrap passwords, the sensitive run variables and any
+    the caller names (a per-host agent token)."""
+    creds, connection = creds or {}, connection or {}
+    values = [creds.get("become_password"), creds.get("vault_password"),
+              connection.get("password"), connection.get("become_password"),
+              *_strings(secret_vars), *(redact or [])]
+    found = {v for v in values if isinstance(v, str) and len(v) >= _MIN_SECRET_LEN}
+    return sorted(found, key=len, reverse=True)  # longest first: no partial masks
+
+
+def scrub(line: str, secrets: list[str]) -> str:
+    for value in secrets:
+        if value in line:
+            line = line.replace(value, MASK)
+    return line
+
+
 async def start_run(
     app: "Quart",
     run_id: str,
@@ -303,6 +342,7 @@ async def start_run(
     inventory_content: str | None = None,
     connection: dict | None = None,
     secret_vars: dict | None = None,
+    redact: list[str] | None = None,
 ) -> None:
     """Execute ansible-playbook as a subprocess and update the DB run row.
 
@@ -315,10 +355,12 @@ async def start_run(
     never persisted on the AnsibleRun row, never placed on argv, and not logged.
     secret_vars are run-time playbook variables flagged sensitive; like
     connection they reach the subprocess (merged into the extra-vars file) but
-    are never persisted.
+    are never persisted. Every value in run_secrets() — redact names more — is
+    masked in each output line before it goes anywhere.
     """
     from steward.models.ansible import AnsibleRunStatus
 
+    secrets = run_secrets(app.config.get("ANSIBLE", {}), connection, secret_vars, redact)
     _run_lines[run_id] = []
     db_output = ""
     truncated = False
@@ -432,7 +474,7 @@ async def start_run(
             assert proc.stdout is not None
 
             async for raw_line in proc.stdout:
-                line = raw_line.decode(errors="replace").rstrip("\n\r")
+                line = scrub(raw_line.decode(errors="replace").rstrip("\n\r"), secrets)
                 _broadcast(run_id, line)
                 if logf is not None:
                     logf.write(line + "\n")
@@ -478,7 +520,7 @@ async def start_run(
         final_status = "cancelled" if run_id in _cancelled else "failed"
         # Surface WHY in the run output instead of an empty "failed" (e.g. ENOSPC
         # when the temp dir can't be created — the run never reaches Ansible).
-        msg = f"[run error] {type(exc).__name__}: {exc}"
+        msg = scrub(f"[run error] {type(exc).__name__}: {exc}", secrets)
         _broadcast(run_id, msg)
         if len(failures) < _FAILURE_CAP:
             failures.append(msg[:500])