BLACK_WALLhomeEN · ES · PT
FAILURE_MODE_TAXONOMY

The 28 ways an AI agent breaks things — and the flags that catch them.

Black_Wall checks every agent action against these 28 documented failure modes before it runs. Each is a real red-flag code returned by the forecast API.

Financial

AMOUNT_UNVERIFIED
Paying an amount the agent never verified
An agent moves money using an amount pulled from context without re-checking it against the invoice or source of truth —
AMOUNT_OUT_OF_BAND
An agent payment far larger than normal
The payment amount is wildly out of line with what this agent or recipient usually transacts — a classic signal of a mis
CURRENCY_MISMATCH
An agent paying in the wrong currency
The currency in the payload doesn’t match what the recipient expects, causing over- or underpayment or a failed settleme

Communications

RECIPIENT_UNVERIFIED
Sending to an unverified recipient
The recipient address or identifier hasn’t been validated recently — a typo or stale contact sends sensitive content to
TONE_MISMATCH
An agent message with the wrong tone
The message tone differs sharply from prior communication with this recipient — often a templating error, wrong context,
TIMING_SUSPICIOUS
An agent sending at a suspicious time or rate
Sending at an unusual hour or in a burst far above the normal pattern — frequently a runaway loop or a compromised job.
PII_EXPOSURE
An agent leaking personal data (PII)
The message contains personally identifiable information that may not be authorized to send — a privacy and compliance b
LEGAL_LANGUAGE_DETECTED
An agent making legal commitments
The message contains commitments or legalese that should require human review before it is sent.
MASS_RECIPIENT
An agent mass-emailing too many recipients
The recipient list exceeds a safe threshold (default >50) — one bad send reaches thousands and can’t be recalled.

Data mutation

IRREVERSIBLE_NO_BACKUP
An irreversible action with no backup
The action cannot be undone and no backup or snapshot exists — if it’s wrong, the data is simply gone.
SQL_NO_WHERE
An agent ran UPDATE or DELETE with no WHERE clause
A DELETE or UPDATE statement without a WHERE clause applies to every row in the table instead of the intended one.
DESTRUCTIVE_VERB
An agent ran a destructive command (DROP / TRUNCATE)
The statement uses DROP, TRUNCATE, FORCE DELETE, or similar — instantly and irreversibly destructive.
CROSS_ENVIRONMENT
An agent hitting production from a non-prod context
The action appears to target production when the intent was staging or test (or vice versa) — the right command on the w
SCHEMA_DRIFT
An agent assuming a stale database schema
The action assumes a schema that may have changed since the agent last saw it — corrupting data or failing mid-write.

Authorization

MISSING_AUTH
An agent acting without required approval
Required authorization or approval is not evidenced in context for a sensitive action.
PERMISSION_ESCALATION
An agent granting itself more access
The action would grant access beyond what the agent typically holds — widening the blast radius of any future mistake.
KEY_ROTATION_ANOMALY
An agent rotating or revoking keys unexpectedly
Rotating or revoking keys without the normal preconditions can lock out systems or users.

Intent integrity

AMBIGUOUS_INTENT
An agent acting on unclear intent
The stated user intent is unclear, contradictory, or missing — so the agent is guessing at a consequential action.
INTENT_MISMATCH
The action doesn’t match what was asked
The action diverges from the stated user intent — a sign the agent drifted off-task or was steered elsewhere.
REGEX_OVERFITS_INPUT
An agent acting on fragile pattern-matched input
The inputs look like they were lifted via a brittle pattern match, so they’re likely the wrong values.
PROMPT_INJECTION_LIKELY
Prompt injection in an agent’s tool inputs
The inputs contain language characteristic of prompt injection — an attacker steering the agent’s real-world action thro

Operational

RATE_ANOMALY
An agent firing an action far too often
The agent is performing this action much more frequently than its baseline — usually a runaway loop.
RETRY_LOOP_RISK
An agent retrying a failed action with no backoff
It looks like a retry of a recently-failed action with no backoff — risking duplicates, double-charges, or a cascade.
EXTERNAL_API_UNRELIABLE
An agent relying on an unstable external API
The action depends on an external API with recent instability or undocumented behavior — so the data driving it may be w
COST_ANOMALY
An agent action that burns abnormal resources
The action would consume far more resources, tokens, or spend than is typical.
STALE_CONTEXT
An agent acting on stale context
The context references events older than 24 hours that may have changed — the agent is acting on outdated facts.

Compliance

REGULATORY_RISK
An agent action touching a regulated domain
The action touches a regulated domain (PCI, HIPAA, SOX, GDPR) with no evidence of a compliance check.
POLICY_DRIFT
An agent action that violates company policy
The action would violate a stated company policy established earlier in context.

Try it on your own action

No signup — paste an action and see the verdict.

Try it live →Get a free API key