AgentPROV corpora: AgentDojo-format studies II (MIT)

Execution-provenance corpora built by AgentPROV v0.1.0 from third-party AgentDojo-format run logs published under the MIT licence (MELON variants, defence studies and injection experiments): each execution as an AgentPROV Execution Record, semantic annotations and W3C PROV bundles, validated and hash-chained. This deposit draft holds 7 AgentPROV corpus version(s), prepared offline by agentprov package --deposit for the release date 2026-10-30. Nothing has been uploaded or published. Corpus kongminos-agentdojo 0.1.0: jspace-agentdojo-950: AgentDojo 0.1.35 tool_knowledge logs of DeepSeek V4-Flash (kongminOS/jspace-agentdojo-950). 950 execution records, of which 0 are outcome-only records. Licence class: MIT. Content digest sha256:c0f6f25257b409b7dabf7055432af89b66af27aa2910453b0b9bb70a32487804; root digest sha256:8daff8e33e923dd808c436968ee5d2a681747eaad10d9d2f1c314a5d58344434. Corpus linyxus-agentdojo 0.1.0: Linyxus/agentdojo lacuna-exp: AgentDojo 0.1.35 logs of the Tacit defence on OpenRouter models. 9068 execution records, of which 3 are outcome-only records. Licence class: MIT. Content digest sha256:e9372d40292cb5a9fdc94620d85c0fde8d050c048ce4831c4048b50d03deb390; root digest sha256:2cf616e78ba114e3f3d3101c714e8a7d327146929de0937e1a53b198bafa4b34. Corpus melon-agentdojo 0.1.0: MELON variants on AgentDojo: AgentDojo 0.1.35 logs of masked re-execution defences (xwsdid/MELON). 1911 execution records, of which 0 are outcome-only records. Licence class: MIT. Content digest sha256:b454f0d8f4589efdb31c32d0236cd2a3a78c8bdb8aae3c0e2f8eb1a183ff9615; root digest sha256:dab02b533eb04e4b247f8bf499acb0e1685a892aa8ec3ddf96eefdac3b724cab. Corpus memo-ozdincer-agentdojo 0.1.0: memo-ozdincer/agentdojo results containers: circuit-breaker (LoRRA) adapters and AgentDojo defences on Llama-3.3-70B and Qwen2.5-7B. 13064 execution records, of which 1038 are outcome-only records. Licence class: MIT. Content digest sha256:41f0810c48c4670faa8bc85b00500c3c65898d9d5894e457cb7aa9f5c6fea7c0; root digest sha256:bac49167e2ad25d4f65a76c9d91919dd7108b8dff6e33d61a9fe1d47c576d0fe. Corpus menelaus29-ipi-multidomain 0.1.0: IPI multidomain agent study: AgentDojo 0.1.35 logs of Gemma 4 and Gemini Flash with a 197-payload corpus (Menelaus29). 858 execution records, of which 1 are outcome-only records. Licence class: MIT. Content digest sha256:48990f53102c781715432aa8c3e94eb2b383de272209035c188521e3b306cdab; root digest sha256:19d246f6f71a228a08760fec6407a4c82738941e85fec8a0f8f4a3d2933cfa6d. Corpus mohameduk-agentdojo 0.1.0: UBAG gateway on AgentDojo: AgentDojo fork logs of an authorization-gateway defence (mohameduk/agentdojo). 3628 execution records, of which 0 are outcome-only records. Licence class: MIT. Content digest sha256:c959d62da1111d604bcb4ba44cc536142893ddae664d5eabbf55b33d4b643d69; root digest sha256:afa8abd130086c303fa252005aa02f2b7ec1349a8286415d82ea5899959e964f. Corpus shreeyavarma-llm-safety 0.1.0: llm_safety: AgentDojo 0.1.34 workspace logs of gpt-4o and local Qwen agents (ShreeyaVarma/llm_safety). 672 execution records, of which 2 are outcome-only records. Licence class: MIT. Content digest sha256:8f9c3e9f206c522d27dc96ab1d433302aae217c3e38a08e1c88320b9932623cb; root digest sha256:f9967d4ec4cf3d867e217ac25b6dc0d7da8d7907ff84a85b005f798908e1c9d1. Counts are computed from each corpus manifest at sealing. Each DATASHEET states what the records do and do not establish; a statement about the data is scoped as it is there. To verify: compare each archive's sha256 or md5 with DEPOSIT-MANIFEST.json, run agentprov verify on the archives (archive mode, no extraction; H.5.5), and compare each root digest with the one published outside the deposit (the corpus tag annotation and the paper).