Papers › No Task Fails Every Time: Why One-Shot Audits Are Structurally Blind to Agent Damage

No Task Fails Every Time: Why One-Shot Audits Are Structurally Blind to Agent Damage

15 Aug 2026arXiv:2608.15286added by Syntology

Shiven Khurdi

Title, abstract, authors and date from arXiv's metadata (CC0); this paper is not in the Papers with Code archive (frozen 2025-07-28).

We introduce AgentRelBench, an environment-agnostic reliability instrument that computes ground-truth, severity-priced damage from database state diffs across repeated runs, with no LLM in the measurement path, demonstrated on EnterpriseOps-Gym. Across 2,128 evaluation runs spanning nine models in six families (four development, three pre-registered held-out, plus a frontier pass on two frontier-tier models that the pre-registration designates exploratory), we find: (1) damage on irreversible actions is universal across the families we measured and stochastic within them on pinned, single-provider stacks. (2) No task damaged on every run: zero always-fail cells across 42 confirmatory held-out damage events. A single clean run misses a damage-producing (model, task) pair 0.80 of the time on the development pool (13 pairs); the held-out pool is descriptively consistent (0.575 over 5 pairs, pair-weighted) but sits below our pre-registered power floor and is reported as underpowered, not as confirmation. (3) Damage-producing task count falls with model capability, from 7 of 20 tasks for an 8B model to 1 of 20 for the most capable; capability is confounded with family and training, so this is an observed gradient, not a causal claim. The residual damage does not change in character: in the exploratory frontier pass, the most capable model's one damaging task damages at p̂ = 0.16 per run, inside the same demonstrably-stochastic band, and a single audit misses it 84% of the time. (4) One model family committed the gated irreversible change while declaring it had refused: transcript- and judge-based grading scores those runs as safe refusals, only state diffs as damage. All confirmatory findings were pre-registered with per-claim demote criteria; one demoted our own initially favored finding, which we report.

Code

Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

11 samples harvested; 9 ran; 0 honoured the contract we drafted; 2 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

Licence: 0 of the 11 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from shivenkk/agentrelbench. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

Results from the paper

The Papers with Code archive ends with its 2025-07-28 snapshot. This paper's arXiv identifier, 2608.15286, was issued in August 2026, after that date, so the archive has no leaderboard rows for it.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections