fix(skills_guard): agent-created dangerous skills ask instead of block (#2446)

fix(skills_guard): agent-created dangerous skills ask instead of block

tests/tools/test_skills_guard.py

@@ -167,12 +167,12 @@ class TestShouldAllowInstall:
         assert allowed is True
         assert "agent-created" in reason
 
-    def test_dangerous_agent_created_blocked(self):
-        """Agent-created skills with dangerous verdict (critical findings) stay blocked."""
+    def test_dangerous_agent_created_asks(self):
+        """Agent-created skills with dangerous verdict return None (ask for confirmation)."""
         f = [Finding("env_exfil_curl", "critical", "exfiltration", "SKILL.md", 1, "curl $TOKEN", "exfiltration")]
         allowed, reason = should_allow_install(self._result("agent-created", "dangerous", f))
-        assert allowed is False
-        assert "Blocked" in reason
+        assert allowed is None
+        assert "Requires confirmation" in reason
 
     def test_force_overrides_dangerous_for_agent_created(self):
         f = [Finding("x", "critical", "c", "f", 1, "m", "d")]

tools/skill_manager_tool.py

@@ -59,9 +59,15 @@ def _security_scan_skill(skill_dir: Path) -> Optional[str]:
     try:
         result = scan_skill(skill_dir, source="agent-created")
         allowed, reason = should_allow_install(result)
-        if not allowed:
+        if allowed is False:
             report = format_scan_report(result)
             return f"Security scan blocked this skill ({reason}):\n{report}"
+        if allowed is None:
+            # "ask" — allow but include the warning so the user sees the findings
+            report = format_scan_report(result)
+            logger.warning("Agent-created skill has security findings: %s", reason)
+            # Don't block — return None to allow, but log the warning
+            return None
     except Exception as e:
         logger.warning("Security scan failed for %s: %s", skill_dir, e, exc_info=True)
     return None

tools/skills_guard.py

@@ -43,7 +43,7 @@ INSTALL_POLICY = {
     "builtin":       ("allow",  "allow",   "allow"),
     "trusted":       ("allow",  "allow",   "block"),
     "community":     ("allow",  "block",   "block"),
-    "agent-created": ("allow",  "allow",   "block"),
+    "agent-created": ("allow",  "allow",   "ask"),
 }
 
 VERDICT_INDEX = {"safe": 0, "caution": 1, "dangerous": 2}
@@ -659,10 +659,17 @@ def should_allow_install(result: ScanResult, force: bool = False) -> Tuple[bool,
 
     if force:
         return True, (
-            f"Force-installed despite blocked {result.verdict} verdict "
+            f"Force-installed despite {result.verdict} verdict "
             f"({len(result.findings)} findings)"
         )
 
+    if decision == "ask":
+        # Return None to signal "needs user confirmation"
+        return None, (
+            f"Requires confirmation ({result.trust_level} source + {result.verdict} verdict, "
+            f"{len(result.findings)} findings)"
+        )
+
     return False, (
         f"Blocked ({result.trust_level} source + {result.verdict} verdict, "
         f"{len(result.findings)} findings). Use --force to override."
@@ -694,7 +701,12 @@ def format_scan_report(result: ScanResult) -> str:
         lines.append("")
 
     allowed, reason = should_allow_install(result)
-    status = "ALLOWED" if allowed else "BLOCKED"
+    if allowed is True:
+        status = "ALLOWED"
+    elif allowed is None:
+        status = "NEEDS CONFIRMATION"
+    else:
+        status = "BLOCKED"
     lines.append(f"Decision: {status} — {reason}")
 
     return "\n".join(lines)