{"version":"6.3.0","api_prefix":"/gradio_api","mode":"blocks","app_id":15490048520663558959,"dev_mode":false,"vibe_mode":false,"analytics_enabled":true,"components":[{"id":1,"type":"dropdown","props":{"choices":[["Start Here","Start Here"],["Why verification matters","Why verification matters"]],"value":"Start Here","type":"value","allow_custom_value":false,"filterable":true,"label":"Pages","show_label":true,"container":true,"min_width":160,"visible":true,"elem_classes":[],"preserved_by_key":["value"],"buttons":[],"name":"dropdown","_selectable":false},"skip_api":false,"component_class_id":"ab4364d7fce4d39b423da06814d2184031a9485af4e94c2e684a84bfb6e53489","key":null,"api_info":{"type":"string","enum":["Start Here","Why verification matters"]},"api_info_as_input":{"type":"string","enum":["Start Here","Why verification matters"]},"api_info_as_output":{"type":"string","enum":["Start Here","Why verification matters"]},"example_inputs":"Start Here"},{"id":2,"type":"markdown","props":{"value":"# RFTSystems — Agent Forensics Suite\n**Evidence-first instrumentation for AI agents and safety decisions.**\nAudit, prove, replay, and diff runs — turning “trust me” into verification.\n\n## Why I built this\nAI is being shipped into real systems faster than teams can reliably reproduce or explain agent behaviour. When an agent fails, too many postmortems still rely on screenshots, partial logs, and opinions — not evidence.\n\nThe operational risk is not only that an agent does the wrong thing. The deeper risk is that **nobody can prove what happened**: what the system saw, what it decided, what it called, what it wrote, and where the run diverged. When failures are unreproducible, accountability collapses.\n\nRFTSystems exists to make behaviour **inspectable and independently verifiable**. This suite produces evidence bundles you can share and validate: Ed25519-signed receipts, hash-chained timelines, deterministic replays, Merkle proofs, and first-divergence diffs. You don’t need to trust the author — you can verify the evidence.\n\nI can’t promise “AI will never take over.” No one can. What I *can* promise is this: **with chain-of-custody logs and signed receipts, we can prove what happened and who is responsible.**\n\n## The workflow\n**learn → generate proof → record reality → seal it → replay → diff → audit → benchmark**\n\n### Quick start (60 seconds)\n1. Open **AuditPlane** and generate a baseline suite.\n2. Replay the same suite and confirm drift diffs (should be 0 if unchanged).\n3. Export the offline bundle — anyone can verify receipts and Merkle proofs.\n\n### Agent pipeline (real systems)\n1. **Record reality** (Agent Flight Recorder).\n2. **Seal it** into receipts (RFT Memory Receipt Engine).\n3. **Diff** two runs and find first divergence (TimelineDiff).\n4. **Audit** state transitions and governance evidence (TrustStack).\n5. **Benchmark** verifiable performance signals (Coherent Compute Engine).\n\n## The labs\n- **[AuditPlane — LLM Decision Proofs](https://huggingface.co/spaces/RFTSystems/AuditPlane__LLM_Decision_Proofs)** — Signed verification plane: Ed25519-signed decision receipts + hash-chained runs + replay + drift diffs + Merkle proofs.\n- **[ReplayProof Agent POV Verified Replay](https://huggingface.co/spaces/RFTSystems/ReplayProof__Agent_POV__Verified_Replay)** — Fast proof: generate a deterministic run bundle you can verify and replay anywhere.\n- **[Agent Flight Recorder](https://huggingface.co/spaces/RFTSystems/Agent_Flight_Recorder)** — Chain-of-custody logging: hash-chained events across prompts, tools, outputs, and memory reads/writes.\n- **[RFT Memory Receipt Engine](https://huggingface.co/spaces/RFTSystems/RFT_Memory_Receipt_Engine)** — Proof layer: generate/download tamper-evident receipts; upload to independently verify integrity.\n- **[TimelineDiff Differential Reproducibility](https://huggingface.co/spaces/RFTSystems/TimelineDiff__Differential_Reproducibility)** — First divergence: align two run bundles and pinpoint exactly where/why they split.\n- **[TrustStack Console](https://huggingface.co/spaces/RFTSystems/TrustStack_Console)** — Audit cockpit: inspect runs, compare state, and trace exactly what changed and why.\n- **[Coherent Compute Engine](https://huggingface.co/spaces/RFTSystems/Coherent_Compute_Engine)** — Verification-first benchmark: live throughput + stability/energy behaviour + downloadable receipt.\n\n## Design principle\nWe don’t ‘hand-wave’ agent safety. We measure drift from declared intent and produce evidence. Enforcement remains an operator decision; this suite is the instrumentation layer.\n\n**Tags:** #Agents #LLMOps #MLOps #AISafety #Reproducibility #Forensics #Security #Governance","show_label":false,"rtl":false,"latex_delimiters":[{"left":"$$","right":"$$","display":true}],"visible":true,"elem_classes":[],"preserved_by_key":["value"],"sanitize_html":true,"line_breaks":false,"header_links":false,"container":false,"padding":false,"name":"markdown","_selectable":false},"skip_api":false,"component_class_id":"927e99ba5dda4bcd0dfe596d4d0cecdc2e0b41e7e95071a05f96ae3971390ed6","key":null,"api_info":{"type":"string"},"api_info_as_input":{"type":"string"},"api_info_as_output":{"type":"string"},"example_inputs":"# Hello!"},{"id":3,"type":"accordion","props":{"label":"Licence / Rights Notice (click to expand)","open":false,"visible":true,"preserved_by_key":[],"name":"accordion"},"skip_api":true,"component_class_id":"f7ca9b901f75c307c1f5ddd33f789d906d98b7f72e2885efbfaa76243b6822af","key":null},{"id":4,"type":"markdown","props":{"value":"All materials contained in or associated with this repository — including but not limited to text, code, algorithms, equations, figures, datasets, and documentation — are original works authored by Liam Grinstead and form part of the Rendered Frame Theory (RFT) research framework.\n\nThese works are protected under the following laws and treaties:\n\n• Copyright, Designs and Patents Act 1988 (UK) — ss.1–103 (copyright subsistence, ownership, and infringement) and ss.77–89 (moral rights).\n• Trade Secrets (Enforcement etc.) Regulations 2018 (UK) — Regs.2–6 (protection of confidential know-how, algorithms, and unpublished research).\n• Copyright and Rights in Databases Regulations 1997 (UK) — Regs.14–24 (protection of compiled datasets).\n• Berne Convention for the Protection of Literary and Artistic Works (1886) — Arts.5(2) & 6bis (automatic international copyright and moral rights).\n• TRIPS Agreement (1994) — Arts.9–14 (international enforcement of copyright and related rights).\n\nAll rights are reserved.\n\nNo part of this work may be copied, reproduced, distributed, performed, displayed, trained upon by AI systems, reverse-engineered, or used to create derivative works without the author’s explicit written consent.\n\nEnforcement rights: Unauthorised use constitutes infringement under CDPA 1988 ss.16 & 96–103, giving rise to civil remedies (injunctions, damages, delivery-up, account of profits, and costs recovery).\nCommercial infringement may amount to a criminal offence under CDPA s.107, punishable by fines and/or imprisonment.\n\nVerification: Each record is timestamped through the Zenodo/DataCite registry and may reference the master DOI: https://doi.org/10.5281/zenodo.17460107 as the consolidated legal and authorship archive.\n\n© 2025 Liam Grinstead — All Rights Reserved.","show_label":false,"rtl":false,"latex_delimiters":[{"left":"$$","right":"$$","display":true}],"visible":true,"elem_classes":[],"preserved_by_key":["value"],"sanitize_html":true,"line_breaks":false,"header_links":false,"container":false,"padding":false,"name":"markdown","_selectable":false},"skip_api":false,"component_class_id":"927e99ba5dda4bcd0dfe596d4d0cecdc2e0b41e7e95071a05f96ae3971390ed6","key":null,"api_info":{"type":"string"},"api_info_as_input":{"type":"string"},"api_info_as_output":{"type":"string"},"example_inputs":"# Hello!"},{"id":5,"type":"form","props":{"scale":0,"min_width":0,"preserved_by_key":[],"name":"form"},"skip_api":true,"component_class_id":"ed5cd51f569f3a7624b1709c4e7e50a5a608037f5fb82307671a5cec77e36b71","key":null}],"css":null,"connect_heartbeat":false,"js":null,"head":null,"title":"RFTSystems — Agent Forensics Suite","space_id":"RFTSystems/START_HERE__Agent_Forensics_Suite","enable_queue":true,"show_error":false,"footer_links":["api","gradio","settings"],"is_colab":false,"max_file_size":null,"stylesheets":["https://fonts.googleapis.com/css2?family=Source+Sans+Pro:wght@400;600&display=swap"],"theme":"default","protocol":"sse_v3","body_css":{"body_background_fill":"white","body_text_color":"#27272a","body_background_fill_dark":"#0f0f11","body_text_color_dark":"#f4f4f5"},"fill_height":false,"fill_width":false,"theme_hash":"2fa39f6df94156f7ef9d84370dbb2657d672827c75f082ac995879c32290c5ca","pwa":true,"pages":[["","Home",true]],"page":{"":{"layout":{"id":0,"children":[{"id":5,"children":[{"id":1}]},{"id":2},{"id":3,"children":[{"id":4}]}]},"components":[1,2,3,4,5],"dependencies":[0]}},"mcp_server":false,"i18n_translations":null,"dependencies":[{"id":0,"targets":[[1,"change"]],"inputs":[1],"outputs":[2],"backend_fn":true,"js":null,"queue":true,"api_name":"render_doc","api_description":null,"scroll_to_output":false,"show_progress":"full","show_progress_on":null,"batch":false,"max_batch_size":4,"cancels":[],"types":{"generator":false,"cancel":false},"collects_event_data":false,"trigger_after":null,"trigger_only_on_success":false,"trigger_only_on_failure":false,"trigger_mode":"always_last","api_visibility":"public","rendered_in":null,"render_id":null,"connection":"sse","time_limit":null,"stream_every":0.5,"event_specific_args":null,"component_prop_inputs":[],"js_implementation":null}],"layout":{"id":0,"children":[{"id":5,"children":[{"id":1}]},{"id":2},{"id":3,"children":[{"id":4}]}]},"username":null,"root":"https://rftsystems-start-here-agent-forensics-suite.hf.space"}