Misleading-Cue Red Team¶
Test / assessment — instantiates Predictive-Cue Wayfinding Design
An adversarial exercise that hunts for cues which attract traversal while concealing low relevance, hidden cost, or risk — approaching the interface as an attacker exploiting the gap between attention and truth.
Most wayfinding checks ask whether cues help. The Misleading-Cue Red Team asks the inverse and asks it adversarially: where does a cue win attention while hiding the truth? It is an offense-oriented exercise in which reviewers deliberately try to find — or, in the strong form, to construct — cues that pull an agent down a path whose real relevance, cost, or risk is worse than the cue implies. The defining idea is the attacker's stance: the red team assumes cues can be enticing and false at the same time, and its whole job is to surface that gap before real users or regulators do. It hunts deception, not honest error; it does not neutrally watch users succeed.
Example¶
A personal-budgeting app is preparing to launch a paid tier, and the growth team has been iterating on the upgrade flow. A red team is chartered to attack it. Adopting the mindset of a user who only wants the free features, they walk the flow looking for cues that would route that user into paying without a clear, honest signal. They find several. A prominent "Continue" button on the free-account setup screen actually enrolls the user in a 7-day trial that auto-converts to an annual plan — the button's label promises progress, its destination is a commitment. A "Recommended" badge sits on the priciest plan with no basis in the user's usage. A greyed-out "Maybe later" link is styled to look disabled, discouraging the exit the user is entitled to.
Each finding goes into an exception log with the exploited intent, the concealed cost, and a severity. The auto-converting "Continue" is flagged critical: it hides a recurring charge behind a cue that reads as a neutral next step. The red team does not redesign the flow — it produces the evidence that these cues are attractive and misleading, which is what forces the fix and what the team can be held to.
How it works¶
The distinguishing mechanics are adversarial and deception-focused:
- Attacker's intent model. The team role-plays specific exploited intents ("I only want the free tier," "I'm skimming, not reading") and probes where a cue betrays that intent.
- Attractiveness-versus-truth split. For each suspect cue, they separate what draws the click from what the destination actually delivers — the wider the gap, the higher the severity.
- Hidden-cost hunting. They look specifically for the cost, risk, or eligibility catch that the cue omits — the part a user pays only after committing.
- Exception logging. Every confirmed false cue is recorded with its mechanism of deception, so patterns (not just instances) become visible and the guardrail can be tightened.
Tuning parameters¶
- Adversary aggressiveness — from "flag anything a skimmer could misread" to "only flag deliberate concealment." Aggressive settings catch more but flag honest persuasion as harm.
- Intent personas targeted — which exploited intents the team role-plays; each persona exposes a different class of trap.
- Scope — the whole surface versus known-risky flows (checkout, cancellation, consent). Focused red-teaming goes deeper where harm concentrates.
- Severity rubric — how concealed cost and risk translate to priority; a good rubric ties severity to real user harm, not aesthetic offense.
- Remediation loop — whether findings merely log or block release. Gating on critical findings gives the exercise teeth.
When it helps, and when it misleads¶
The red team earns its place wherever incentives tempt a team to trade user trust for a metric — checkout, consent, cancellation, upsell. It catches dark patterns, the interface tricks Harry Brignull named, before they reach users or draw a regulator's attention.[1] Its adversarial framing surfaces concealment that a cooperative review, focused on whether cues help, structurally cannot see.
Its own failure mode is over-zeal: an aggressive red team can brand every act of persuasion a dark pattern, blurring the line between a genuinely helpful default and a concealed one, and eroding its credibility so real findings get dismissed. It can also become a checkbox that teams learn to game, or fixate on cosmetic offenses while missing structural ones. The guarding discipline is to tie every finding to concrete user harm — a hidden cost, a foreclosed choice, a false promise — and to distinguish legitimate persuasion from concealment rather than treating all influence as attack.
How it implements the components¶
misleading_scent_guardrail— the red team is the active test of this guardrail: it probes for cues that breach the honesty boundary and thereby defines where the guardrail must hold.false_scent_exception_log— each confirmed misleading cue is recorded with its deception mechanism and severity, building the running log of exceptions.traversal_cost_budget— its signature move is exposing the hidden cost or risk a cue conceals, making the true traversal cost explicit.
It hunts deception adversarially but does not neutrally observe representative agents completing real tasks (agent_intent_context, decision_point_inventory, wayfinding_feedback_signal) — that cooperative, observational measurement is Task-Based Wayfinding Test.
Related¶
- Instantiates: Predictive-Cue Wayfinding Design — supplies the adversarial check that keeps cues honest by attacking them.
- Sibling mechanisms: Breadcrumb and Landmark Trail · Cue-Destination Alignment Matrix · Destination Preview Card · Link-Label Scent Audit · Progressive Disclosure Preview · Route Recovery Pattern · Scent Clickthrough Trace Dashboard · Task-Based Wayfinding Test
Editorial Notes¶
Form Classification¶
Form family: Experiment, Test & Rehearsal
Rationale: Misleading-Cue Red Team operates as a bounded trial, probe, simulation, or rehearsal that generates evidence from performance because it an adversarial exercise that hunts for cues which attract traversal while concealing low relevance, hidden cost, or risk — approaching the interface as an attacker exploiting the gap between attention and truth.
Independent corroboration: The frozen evidence defines Misleading-Cue Red Team as 'An adversarial exercise that hunts for cues which attract traversal while concealing low relevance, hidden cost, or risk — approaching the interface as an attacker exploiting the gap between attention and truth', so its operative form is Experiment, Test & Rehearsal.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Human-Computer Interaction
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Adversarially testing wayfinding cues for deceptive affordances belongs to HCI and usability evaluation.
Related originating lineages:
- Psychology — Attention and cue-validity research explains why salient signals can mislead traversal.
- Security Studies & Intelligence Analysis — Red-team practice supplies the attacker mindset and exploit search.
Review resolution: Both independent reviews place the primary provenance in human_computer_interaction. The queued differences (domain_reach_disagreement) concern secondary metadata, not primary lineage. The final retains psychology, security_intelligence only where a reviewer supplied a formative-lineage rationale; downstream use or broad applicability by itself is not treated as origin. origin_mode=cross_disciplinary_synthesis because the supplied rationales identify formative contributions that are composed in the mechanism's present form. domain_reach=multi_domain records established application breadth separately from provenance. confidence=high preserves the more cautious evidence assessment. encyclopedia_synthesis=true records whether either reviewer identified deliberate corpus-level composition.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
The red team overlaps the Link-Label Scent Audit in surface but not in stance: the audit assumes cues are trying to be honest and looks for drift and error; the red team assumes cues may be engineered to mislead and looks for concealment. A space usually needs both — the audit for the slow decay of well-meant labels, the red team for the deliberate trap.
References¶
[1] Dark patterns — a term coined by Harry Brignull in 2010 — are interface choices designed to trick users into doing things they would not otherwise do, such as hidden costs, disguised ads, or obstructed cancellation. The red team's target set is essentially the wayfinding subset of these: cues attractive by design and misleading by consequence. withdrawn registry ↩