How about the following powers:
Hellish Rebuke - If you take damage before the end of your next turn, the target takes an extra 1d6 + Constitution modifier fire damage.
Riposte Strike - If the target attacks you before the start of your next turn, you make your riposte against the target as an immediate interrupt: a Strength vs. AC attack that deals 1[W] + Strength modifier damage.
Is your feeling that the target should have no way to notice the effects of such powers? Is there no change in the target or attacker, in terms of stance or readiness, lingering flames? Do the powers fail to work if they are a deterrent against further attack, rather than a hidden gotcha of the attack?
See, those are good examples ;-).
FWIW, I agree that Brash Assault is terrible, and don't see that Hellish Rebuke or Riposte Strike need to not signal what they're doing.
My feeling is that they both don't
need to be traps, but conversely, neither do the
need to be known. Let's look into a little more detail.
Both of these powers are weak if the target never had any intention of doing the triggering action in the first place; and if they never had such intention, then it's irrelevant whether the target knows or not.
So, to compensate for the weakness when used in a case in which the trigger wasn't going to go off anyhow, it should be a little stronger in the case that (assuming no foreknowledge) the trigger
was going to fire.
In the case of the Hellish Rebuke,
other enemies certainly don't know of the effect automatically, so the difference isn't actually very large. After the first time it triggers, I assume the effect is understood, so then the difference isn't large either. If the creature intends to trigger despite knowing what it does, there's no difference. So the difference is very minor: only
if the creature was intending to do the triggering action and
that creature was the first to trigger it, and
if the creature has a better choice if it knows, and
if this is the first time it would trigger, and
if it actually succeeds to trigger (not on an attack, but on damage) is the difference significant. And the extent of that significance is (worst case) 1/enc 1d6+mods difference.
The story with Riposte strike is different, but has a similar global conclusion: whether the monster knows or not isn't actually a huge deal. In particular, it targets AC, so the target really needs a significant leg up to beat piercing strike.
So in both cases, I think either interpretation is OK. Really fully using Hellish Rebuke probably requires getting damaged in other ways (ongoing damage, zones, weak friendly fire, enticing other creatures to damage you etc.), and if you're doing that, the knowledge question is moot; the trigger will fire fairly quickly and people will understand by virtue of being affected then.
Finally, note that we're talking about the minimal rules-based knowledge creatures get
by default. For example, I find it fairly reasonable to assume that creature can see threatening reach; the DMG even gives an example with a fire-aura being noticeable before being affected. So it's not like the DM is
forced to withhold knowledge, just that it's not
automatic. Knowledge is granted automatically when you're affected (which under this interpretation would exclude the triggers in these powers) but it's
also granted whenever it should be reasonably visible.
As to in-game logic, particularly with Hellish rebuke, I find quite plausible that some lingering effect might be felt, but a little less plausible that the exact detail (trigger: when he gets damaged) and exact consequence (1d6+mods) is obvious. So, since the balance is largely moot, I'd pick the interpretation that the target doesn't know since that seems to me to make more in-game sense: he may feel a lingering fire, but not exactly what that fire does; at least until it triggers once.
Well, I
would if it'd ever come up ;-). Which it certainly hasn't recently.