ONE PULL REQUEST · ONE REASSESSMENT
See what the merge
evidence actually proves.
An inspectable evidence pack for one supported public GitHub pull request. Git history, CI on the relevant commit, and human approval bound to the candidate.
Inspect the Kiota methodology sample ↗Public sample of the original local Git checks. Microsoft is not a customer; no endorsement is implied. The sample does not include this version’s CI and approval checks.
The G1 study
We analyzed 2037 merged, agent-authored pull requests across 32 public repositories. 24.1% triggered one or more evidence-gap conditions under the checks merge-proof implements.
Read that sentence precisely. It is a statement about missing evidence, not about defects.
This study does NOT claim that 24.1% of agent-authored PRs contained bugs, were broken, were unsafe, or should not have merged. It makes no claim about the correctness of any pull request in the corpus.
- 1547 VERIFIED (75.9%); 490 NOT_PROVEN (24.1%); 0 FAIL. 42 unresolvable PRs were excluded from all rates.
- BASE_DRIFT_UNVERIFIED: 410 PRs (20.1%). PROTECTED_BOUNDARY: 109 PRs (5.4%). A PR can trigger both.
- Only 2 of 6 conceptual checks were run. CI_RAN_ON_FINAL_HEAD, HUMAN_APPROVAL_PRESENT, CANDIDATE_DURABLE_ON_REMOTE and SCOPE_CREEP_VS_DECLARED_SCOPE were not implemented and not measured in this study.
- This is a purposive sample, not a random one. These rates should not be read as rates for GitHub as a whole. All 32 contributing repositories landed in the team stratum; the small/solo stratum is empty.
- The historical study detects agent co-author trailers in squash merges, with a per-repository cap of 80 PRs. Rebase-merged PRs are missed entirely. This differs from the supported open/two-parent history of the Standard Evidence Pack.
VERIFIED
Implemented checks found no blocking evidence gap. It does not prove correctness.
NOT_PROVEN
Evidence is missing under implemented checks. It does not mean defective code.
FAIL
The verifier could not safely establish a result. Required prerequisites need attention. It does not mean bad code.
What you receive
A PDF report, JSON results, capture and evidence manifest, Action handoff Markdown, and one reassessment with a before-and-after diff. Your original report stays intact.
Download securely from this page. Access and one reassessment are available for seven days after payment. Save your private access link and download the pack before it expires.
No call required. No customer source code is executed.
Check eligibility, save your private link, and pay through Stripe. Return here to authorize the public repository and confirm the exact scope. This page shows progress and provides your downloads. Within seven days, prepare your one reassessment of the same PR, confirm its current scope, and download the comparison and Action handoff.
One configured price
Loading Stripe price…
Check eligibility before payment. Stripe confirms the actual transaction amount. A missing-evidence verdict is a delivered result, not a failed run.
If delivery fails, retry or request an exception. Undeliverable purchases are escalated to Ryan as REFUND_REQUIRED. Refunds are reviewed by Ryan, never issued automatically.
The limits are part of the result.
One open PR or reconstructable two-parent merge on public GitHub. Full history must fit the bounded runner: repository metadata ≤128 MiB, fetched objects ≤192 MiB, ≤100,000 reachable commits, and a two-minute run limit. Squash/rebase merges, private repositories, and ambiguous history are not supported in this version.
You confirm the required CI check names and their GitHub App IDs. This pack checks those requirements and one non-author human collaborator approval. It does not certify all branch rules, CODEOWNERS, test quality, or the eventual merge state of an open PR.
This is not bug detection, a security or compliance certification, proof of correctness, or a guarantee of safe code or a safe merge.
These remain outside the implemented proof:
-
CANDIDATE_DURABLE_ON_REMOTE— whether the candidate remains durable on the remote. -
SCOPE_CREEP_VS_DECLARED_SCOPE— whether changes stay within declared scope.
Your evidence pack
START WITH ELIGIBILITYAnyone with this link can access your pack until expiry. Keep it private. It is not sent to Stripe or GitHub.
Your repository is eligible. Continue to the configured Stripe Payment Link, then return here after payment.
Confirm the exact scope
The report applies to these captured states and the listed required checks. If the candidate changes before execution, it must be confirmed again.