FA-99786 / Tool call schema validation / Open access

Invalid array elements escape item validation · case 01

Invalid array elements escape item validation.

Verified by executionVariant 1 · 6 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

Only the first array member is checked.

THE FAILURE

Only the first array member is checked.

Unsuccessful approach: Rejecting arrays longer than one fixes some examples but excludes valid arrays.

Case contract

Every array item must be in the declared allowed set; empty arrays are valid.

Why this case matters

Tool argument validation prevents malformed structured payloads from entering a deterministic downstream operation.

1 / The failure

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(values, allowed):
    return not values or values[0] in allowed
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('valid singleton',solve([N],[N]),True)
check('invalid first',solve(['x',N],[N]),False)
check('invalid later',solve([N,'x'],[N]),False)
check('empty array',solve([],[]),True)
check('all allowed repeats',solve([N,N],[N]),True)
check('all disallowed',solve(['x','y'],[N]),False)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

Boundary fixture	Actual	Expected	Outcome
valid singleton	`True`	`True`	Passed
invalid first	`False`	`False`	Passed
invalid later	`True`	`False`	Failed
empty array	`True`	`True`	Passed
all allowed repeats	`True`	`True`	Passed
all disallowed	`False`	`False`	Passed

SHA-256 / cb6aebafa77f7c5d08603a76141d152e148e0abcc01fb90bdd92a3f7c4e86237

2 / The unsuccessful fix

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(values, allowed):
    return len(values)<=1 and all(v in allowed for v in values)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('valid singleton',solve([N],[N]),True)
check('invalid first',solve(['x',N],[N]),False)
check('invalid later',solve([N,'x'],[N]),False)
check('empty array',solve([],[]),True)
check('all allowed repeats',solve([N,N],[N]),True)
check('all disallowed',solve(['x','y'],[N]),False)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

Boundary fixture	Actual	Expected	Outcome
valid singleton	`True`	`True`	Passed
invalid first	`False`	`False`	Passed
invalid later	`False`	`False`	Passed
empty array	`True`	`True`	Passed
all allowed repeats	`False`	`True`	Failed
all disallowed	`False`	`False`	Passed

SHA-256 / 4e3e26fefbf21d4ed3627aad0cb8de78da366159c1935ed325f9092eaacb1ea9

HELD IN THE MEMBER ARCHIVE

The verified repair and its recorded checks are member-only.

This mechanism has 6 recorded checks per implementation. The open-access tier publishes the failure and the unsuccessful fix; the repaired source that passes every check, and the observations that prove it, are available to members.

Every case sharing this mechanism uses the same contract and the same repair, so this one record is held back for all of them.

Member access is invitation-based. Sign in with your invited account to inspect the repair.

Verification & scope

Self-contained toy JSON-like schema contract over in-memory Python values; no compatibility with a production schema standard or callable binding is claimed. This reproducer isolates one failure mechanism. Results cover the supplied fixtures. Variants within a family share a test contract and should remain grouped when constructing evaluation splits. Related mechanisms with a shared evaluation_group must also remain together; these controlled models are not independent production incidents.

Observations recorded using Python 3.12.14 at 2026-09-30T16:48:28.064631+00:00.

Case digest / 19865ca34ea48da13b8762236ce959c947f7477d7d29fd061e646cda331e0a83