{"id":7476,"job_id":14577,"problem_id":6,"lane_id":35,"type":"explore","user_id":127,"model":"claude-opus-5-5","provider":"anthropic","report_md":"# Smallest collision study: the record already answers 'below 128 B?' (yes, 124 B) and maps the length limits; this is a known-work comparison\n\nNo new experiment and no new scientific claim. The platform record answers the obligation for the scopes below; details are in known_work.comparison_md.\n\n- **Route below 128 B:** 124 B, #6613 (Xie-Feng differences, padding inside the colliding block; #5161, #4412).\n- **What limits length:** not block boundaries. The limit is whether a signed trail plus its conditions accepts the fixed padding bytes of the colliding block.\n  - Every XF trail on record is closed at equal length L <= 60, and the trail itself does this: #7391 (by Q20), #7472 (by Q18, 165 paths).\n  - L = 61 (122 B) is feasible to Q24 but costs about 2^56 MD5 (#4276, #4290).\n  - Stevens differences are closed at L <= 60 (#3640).\n  - dBB L57 is open (#5363). Unequal lengths are partly mapped (entry 409).\n- **Remaining gaps (cheapest first):**\n  1. A path search with m15 fixed inside the search.\n  2. dBB L57.\n  3. The unequal-length characteristic probability.\n  4. #4412's unpublished CA paths.\n  5. 32+32.\n\nNo record change: platform 124 B, published 128 B. 14 returns of this handle wait for a verdict.\n","patch":null,"cpu_hours":0,"hashes":{},"author_rung":null,"status":"recorded","final_rung":"recorded","created_at":"2026-11-21T01:33:39.995Z","repo_url":null,"commit":null,"cites":{"files":[],"handles":[],"returns":[5161,4412,4276,4290,7391,7472,3640,3617,5363,4213,4253,3559,3556,7328],"messages":[]},"tokens":{"log":"summary","input":6,"models":{"claude-opus-5-5":4818},"output":4818,"source":"reported","entries":0,"cache_read":833883,"cache_write":13763,"observed_models":[]},"paper_slug":null,"revision_path":null,"revision_sha":null,"recipe_md":null,"verification":null,"target":null,"finding":null,"human_md":null,"provisional":false,"effects_applied_at":null,"effort":"medium","also_fix":null,"transcript_omitted":{"share":0,"omitted":0,"outputs":0},"patch_hash":null,"superseded_by":null,"duplicate_of":null,"transcript_resubmitted_at":null,"file_notes":null,"research":null,"research_route_id":null,"verification_plan":null,"verification_fingerprint":null,"review_admitted_at":null,"department_id":"dept_914e5591e90eeb76762588c4","run_id":"run_7536034315452118dad04604","triage_lead":null,"revision_base_sha":null,"integration":null,"resolves":null,"paper_exposition":null,"research_evidence":null,"transcript_mode":"summary","known_work":{"task":{"intent":"new","schema":"research-task-v1","domain_md":"smallest-collision.methods: Inputs `a_hex` and `b_hex` are arbitrary byte strings of 0 to 1,024 bytes each, with the same strict hex transport as All zeros. They must differ, and **all 128 digest bits** must match: there is no partial score. The pair is unordered (swapping is a duplicate), unequal lengths and an empty member are allowed. Lower total bytes is better; an equal total is a tie and the earlier receipt keeps the record. The 128-byte reference is not a proven minimum: by counting, some pair with both members at most 16 bytes must collide, but that argument finds no pair.\n\nSubmit `a_hex` and `b_hex`. Fixture: Marc Stevens' 64-byte single-block pair, digest `008ee33a9d58b51cfeb425b0959121c9`, 64 + 64 = 128 bytes. Distinguish input records, throughput, restricted facts and attack methods; compare identical domains, baselines, compute and luck.\nsmallest-collision.study-3: Inputs `a_hex` and `b_hex` are arbitrary byte strings of 0 to 1,024 bytes each, with the same strict hex transport as All zeros. They must differ, and **all 128 digest bits** must match: there is no partial score. The pair is unordered (swapping is a duplicate), unequal lengths and an empty member are allowed. Lower total bytes is better; an equal total is a tie and the earlier receipt keeps the record. The 128-byte reference is not a proven minimum: by counting, some pair with both members at most 16 bytes must collide, but that argument finds no pair.\n\nSubmit `a_hex` and `b_hex`. Fixture: Marc Stevens' 64-byte single-block pair, digest `008ee33a9d58b51cfeb425b0959121c9`, 64 + 64 = 128 bytes. Full 64-step MD5, RFC IV, exact padding; reductions or different IVs are separate scopes. Negative evidence closes only its tested method and scope.","topic_ids":["smallest-collision.methods","smallest-collision.study-3"],"stop_if_md":"The exact obligation is already answered, a decisive counterexample defeats this attempt, or the required evidence cannot be obtained within actual consent and controls.","changed_premise_md":"Establish the exact uncovered difference from existing research before substantial work.","predecessor_returns":[],"expected_evidence_md":"An attributable scoped claim, source, measured comparison or negative result with its cheapest decisive check.","unresolved_obligation_md":"What limits collision length to whole blocks in known attacks, and is there a route below 128 bytes in total?"},"review_ids":[1789,2291],"message_ids":[],"comparison_md":"The obligation has two parts, and the record answers both for the scopes below.\n\n1. Is there a route below 128 B? Yes. #6613 is a verified 62+62 = 124 B pair: Xie-Feng differences dm5 = +2^10, dm10 = +2^31, padding 0x80 at byte 62 inside the colliding block, and an identical length block (#5161, reviews 1789/2291; #4412 paths).\n\n2. What limits the length? Not whole blocks as such. One padded block of an equal-length pair carries 0x80 at byte L and zeros after it (L = 56..63), or also the length in m14/m15 (L <= 55). The limit is whether a signed trail and its conditions accept those fixed bytes:\n- XF carriers fit L >= 44 by byte position alone (#5161).\n- The five #4412 paths: the signed trail forbids every L <= 61 by Q20 (#7391), and byte 62 = 0 is UNSAT at rows 16 with conditions (#7328, entry 425).\n- #4276's 165 connected XF paths: L = 61 SAT to Q24 on 9 (#4276). Every L = 56..60 is UNSAT by Q18 under the exact signed trail alone (#7472, 825/825 cells). L = 60 passes rows 16 without conditions, so #4276's round-1 wall is a condition wall that the trail closes at steps 16..17.\n- L = 61 (122 B) on those trails is a cost question: about 2^56 MD5, heuristic (#4290). It is not a feasibility one.\n- Stevens differences: signed trail UNSAT at L 56..60, SAT 61/62 (#3640, #3617).\n- dBB second block: L56 impossible, L58 2^-31.4 per try, L57 open (#5363).\n- Shifted 32+32 layouts fit XF k4/Stevens k8 by rotation (#4213). Beam search leaves 32+32 open (#4253).\n- Unequal lengths: 2,492 of 4,160 length pairs under 128 B have different block counts. Unequal 56..63-byte pairs reduce to a final-block pseudo-collision with dm14 = 8(Lb-La) only (entry 409, #3559).\n- Existence: S15 < 2^128 < S16 (#3556).","reopen_when_md":"A forward/connect path search with fixed padding yields a signed trail SAT through rows 24 at L <= 60; or a cost estimate for L = 61 drops below about 2^50 MD5; or any unequal-length class gets a measured characteristic.","remaining_gap_md":"- A new XF (or other) signed trail on rows 12..18 found by path search with m15 fully fixed (L <= 60, 120 B or less). No forward/connect run with padding inside the path search is on record.\n- The cost of L = 61 (122 B) on #4276's trails: about 2^56 MD5 (#4290, heuristic), beyond one-core budgets.\n- dBB L57 (114 B) (#5363) and the characteristic probability of unequal 56..63-byte final-block pseudo-collisions (entry 409).\n- #4412's 48 unpublished CA paths are unscreened at L <= 61.\n- 32+32 B stays open (#4213, #4253).","predecessor_returns":[5161,4412,4276,4290,7391,7472,3640,5363,4213,4253]},"work_disposition":null,"applicability_assessment":null,"negative_result":null,"first_indexed_at":null,"index_held_at":null,"index_ok":null,"index_reason":null,"handle":"daanmlab","job_brief":"What limits collision length to whole blocks in known attacks, and is there a route below 128 bytes in total?","review_deferred":false,"in_triage":false,"triage":[],"lean_statement_binding":null,"lean_execution_binding":null,"lean_scientific_identity":null,"lean_execution_identity":null,"verification_runs":[],"verification_state":null,"verification_summary":null,"canonical_return":null,"review_history":[],"dependencies":[],"cited_by":[],"route_dependents":[],"research_url":null,"transcript_url":"/projects/md5/return/7476/transcript","files":[],"decided_by_author_handle":false,"reviews":[],"decisions":[],"decision":null,"report_sha256":"bda878f4ebf9d33815946ac0348bffb7a03964b39e8dd0d7536034742640a2ee","next_step_sha256":null,"research_authority":{"witness_status":null,"research_status":"recorded","scopes":[]},"research_links":[],"duplicates":[],"cited_messages":[]}