FA-99876 / Agent workflow termination / Open access

Workflow: workflow completes when its first branch finishes · case 01

The run closes while another required branch is still active.

Verified by executionVariant 1 · 7 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

The reducer uses any settled branch rather than requiring every branch settled.

THE FAILURE

The reducer uses any settled branch rather than requiring every branch settled.

Unsuccessful approach: Requiring all branches to be successful rejects explicitly failed but settled branches.

Case contract

Given statuses, return true iff the list is nonempty and every item is complete or failed.

Why this case matters

This toy decision rule isolates one bounded agent-loop completion contract.

1 / The failure

Exit 1

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

N = 1
observations = []
def solve(statuses):
    return any(s in ('complete','failed') for s in statuses)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('all complete', solve(['complete']*N),True)
check('one running', solve(['complete','running']),False)
check('settled failure', solve(['failed','complete']),True)
check('empty branches', solve([]),False)
check('single failure', solve(['failed']),True)
check('single waiting', solve(['waiting']),False)
check('unknown unsettled', solve(['complete','paused']),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
all complete	`True`	`True`	Passed
one running	`True`	`False`	Failed
settled failure	`True`	`True`	Passed
empty branches	`False`	`False`	Passed
single failure	`True`	`True`	Passed
single waiting	`False`	`False`	Passed
unknown unsettled	`True`	`False`	Failed

SHA-256 / f59b9c2fcd601e17d35b09315c7ce24b6339239cdddb82b071f96e0630ad5a09

2 / The unsuccessful fix

Exit 1

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

N = 1
observations = []
def solve(statuses):
    return bool(statuses) and all(s == 'complete' for s in statuses)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('all complete', solve(['complete']*N),True)
check('one running', solve(['complete','running']),False)
check('settled failure', solve(['failed','complete']),True)
check('empty branches', solve([]),False)
check('single failure', solve(['failed']),True)
check('single waiting', solve(['waiting']),False)
check('unknown unsettled', solve(['complete','paused']),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
all complete	`True`	`True`	Passed
one running	`False`	`False`	Passed
settled failure	`False`	`True`	Failed
empty branches	`False`	`False`	Passed
single failure	`False`	`True`	Failed
single waiting	`False`	`False`	Passed
unknown unsettled	`False`	`False`	Passed

SHA-256 / e7c79efb0af7db0ee0b539d2ac188d7b6a4020104df3c110ab0096fd66defa6a

HELD IN THE MEMBER ARCHIVE

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

This mechanism has 7 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

Status labels are interpreted as terminal only for complete and failed. 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 2026-09-30T16:48:29.004465+00:00.

Case digest / afccfbc530baef926645b60e2135e08d7fb20f5d355942fff04f06d5d82ab2b1