Stream: Mirror: Isabelle Users Mailing List

Topic: 回复: [isabelle] Soundness-critical issues in thm.ML


view this post on Zulip Email Gateway (Nov 04 2026 at 17:13):

From: Qiyuan XU <cl-isabelle-users@lists.cam.ac.uk>

Dear Isabelle users

I’ve written a patch for this issue and uploaded it to my conda channel. If this issue affects you, you can install the patched version with:

conda create -n <env-name> -c https://conda.qiyuan.me isabelle-bicompose-rename-fix
conda activate <env-name>
isabelle jedit

(Replace <env-name> with any environment name you like.)
isabelle-bicompose-rename-fix is a separate package built on top of the package isabelle which contains the original, unmodified Isabelle. To revert to the original version at any time, simply runconda remove isabelle-bicompose-rename-fix

Alternatively, if you’d prefer to apply the fix yourself, the patch is below:

--- a/src/Pure/thm.ML
+++ b/src/Pure/thm.ML
@@ -2578,10 +2578,10 @@ fun bicompose_aux opt_ctxt {flatten, match, incremented} (state, (stpairs, Bs, B
     fun newAs (As0, n, dpairs, tpairs) =
       let val (As1, rder') =
         if lifted then

-          (case rename_bvars dpairs tpairs B As0 of
+          (case rename_bvars dpairs (rtpairs @ tpairs) Bi As0 of
             SOME term =>
               let fun proof p = Same.commit (Proofterm.map_proof_terms_same term I) p;

-              in (map (strip_apply (Same.commit term) B) As0, deriv_rule1 I proof rder) end
+              in (map (strip_apply (Same.commit term) Bi) As0, deriv_rule1 I proof rder) end
           | NONE => (As0, rder))
         else (As0, rder);
       in

Best wishes
Qiyuan


发件人: cl-isabelle-users-request@lists.cam.ac.uk <cl-isabelle-users-request@lists.cam.ac.uk> 代表 Makarius <makarius@sketis.net>
发送时间: 2026年10月28日 3:35
收件人: kevin.kappelmann@tum.de <kevin.kappelmann@tum.de>; cl-isabelle-users@lists.cam.ac.uk <cl-isabelle-users@lists.cam.ac.uk>
主题: Re: [isabelle] Soundness-critical issues in thm.ML

[Alert: Non-NTU Email] Be cautious before clicking any link or attachment.
On 27/10/2026 09:38, Kevin Kappelmann (via cl-isabelle-users Mailing List) wrote:

Fabian Huch and I recently discovered some soundness-critical issues (attached
below) in thm.ML with the help of AI agents(*).

Thanks for the concrete example, to substantiate what I said on a realated
thread "MCP resources bounds" (Sep-2026):

On 08/09/2026 17:57, Makarius wrote:

There is indeed a fundamental problem here, because the MCP violates
longstanding assumptions about how the prover works, and is used by a human
user. So any "mitigation" can only be symbolic, not really a proper solution.

At FLoC 2026 in Lisbon, I've had some discussions with users of AI + MCP. My
conclusion: we should gradually make the system more robust [...]

Makarius


CONFIDENTIALITY: This email is intended solely for the person(s) named and may be confidential and/or privileged. If you are not the intended recipient, please delete it, notify us and do not copy, use, or disclose its contents.
Towards a sustainable earth: Print only when necessary. Thank you.


Last updated: Nov 19 2026 at 01:46 UTC