FA-99926 / Agent workflow termination / Open access
Workflow: only successful turns debit the workflow budget · case 01
Failed turns leave the allowance unchanged even though every completed turn consumes budget.
ROOT CAUSE
The ledger subtracts successful outcomes only.
THE FAILURE
The ledger subtracts successful outcomes only.
Unsuccessful approach: Charging only failed turns leaves successful work free.
Case contract
Return max(0, allowance - number of completed outcomes), clamped at zero.
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(allowance, outcomes):
return max(0,allowance-outcomes.count('ok'))
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('all used mixed', solve(N,['ok','error']*N),0)
check('failed turn debits', solve(3,['error']),2)
check('successful turn debits', solve(3,['ok']),2)
check('empty outcomes', solve(N,[]),N)
check('overrun', solve(1,['ok','error']),0)
check('empty allowance', solve(0,[]),0)
check('two outcomes', solve(4,['ok','error']),2)
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 used mixed | 0 | 0 | Passed |
| failed turn debits | 3 | 2 | Failed |
| successful turn debits | 2 | 2 | Passed |
| empty outcomes | 1 | 1 | Passed |
| overrun | 0 | 0 | Passed |
| empty allowance | 0 | 0 | Passed |
| two outcomes | 3 | 2 | Failed |
SHA-256 / a1afd1496ad0591aa812757fab4b60de613b77934e974cc189dfe6aebf0a6f36
2 / The unsuccessful fix
Exit 1"""Failure Map reference implementation. Python standard library only."""
import json
N = 1
observations = []
def solve(allowance, outcomes):
return max(0,allowance-outcomes.count('error'))
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('all used mixed', solve(N,['ok','error']*N),0)
check('failed turn debits', solve(3,['error']),2)
check('successful turn debits', solve(3,['ok']),2)
check('empty outcomes', solve(N,[]),N)
check('overrun', solve(1,['ok','error']),0)
check('empty allowance', solve(0,[]),0)
check('two outcomes', solve(4,['ok','error']),2)
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 used mixed | 0 | 0 | Passed |
| failed turn debits | 2 | 2 | Passed |
| successful turn debits | 3 | 2 | Failed |
| empty outcomes | 1 | 1 | Passed |
| overrun | 0 | 0 | Passed |
| empty allowance | 0 | 0 | Passed |
| two outcomes | 3 | 2 | Failed |
SHA-256 / 95f2d079b86eb5ff1174e4f75c318395f0671df4c01ec75395f61ec7634c3b92
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.
Sign in to the archive ↗Verification & scope
Each item in outcomes is one completed turn, labeled ok or error. 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 2026-09-30T16:48:29.887613+00:00.
Case digest / 1bd663c723ff6047044df579057dc8a0c589618fda7bd1fd24915e0dbd42732b