aide-flagship-twin-mind-train SKILL.md skillA
aide-flagship-twin-mind-train from anonymousnomad/aide-sovereign-workbench is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add anonymousnomad/aide-sovereign-workbench/aide-flagship-twin-mind-train@git:20260909.cac77c1Pins git:20260909.cac77c1 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- anonymousnomad/aide-sovereign-workbench
- Path
skills/packs/aide-flagship-twin-mind-train/SKILL.md- Versions recorded
- 1
- Latest version
- git:20260909.cac77c1, committed Sep 9, 2026
- Size
- 10,406 bytes
- Repository stars
- 53
- License
- Apache-2.0
Continue and productionize the CipherV2 twin-mind MoE flagship pretrain (preset local140a50, ~146.5M total, 48 routed experts = 24 Sheldon + 24 Spock + 2 shared, top_k 8, quorum 2/population) on THIS machine GTX 1060 6GB fp32-only Windows. Use whenever resuming/continuing the flagship pretrain, launching the v6 trainer for local140a50, probing VRAM/throughput, running the 6-check twin-mind battery, or deciding batch/seq/curriculum for the flagship run. Verified-shape commands, dependency…
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# AIDE Flagship Twin-Mind MoE — Pretrain Continuation SOP ## The Mission Train the in-house AIDE model: a **twin-mind MoE** where the routed expert pool is split into two typed populations that are structurally forced to disagree (Sheldon = rule/checklist/adversarial; Spock = probability/synthesis), with a shared consensus pool and a per-token quorum so BOTH minds act on every token. This is the committed, user-approved size lane: **~150M total** (fits the 100-150M hard cap). The architecture already exists and has a parked 300-step checkpoint; the job is probe -> continue -> gate -> SFT -> serve. ## Verified Reality (this machine, 2026-09-10) — DO NOT RE-VERIFY, DO NOT GUESS | Fact | Value | Where | |---|---|---| | Training Python | `E:\felon_workspace\venv_py310\Scripts\python.exe` → torch 2.7.1+cu118, CUDA True | bash check 2026-09-10 | | SFT/convert Python (tokenizers) | `E:\felon_workspace\venv_cipher\Scripts\python.exe` (tokenizers 0.23.1) | AGENT_NOTES 2026-09-08 | | Trainer | `E:\FSI-FELON\models\fsi_felon_cipher\train_v6.py` (presets: local30a12, local140a50, 600a150, 1200a200, sft_local30a12) | train_v6.py | …
Read the whole file at its exact version.
How to install
mdr add anonymousnomad/aide-sovereign-workbench/aide-flagship-twin-mind-train@sha256:b03f0853e4558f0bThe line at the top of this page pins the label git:20260909.cac77c1, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_vbu42jb4jaf3dqcj)
4 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 1 version of this file, committed on Sep 9, 2026.
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (10406 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
anonymousnomad/aide-sovereign-workbench · 53 stars · license Apache-2.0 · pushed 2026-10-26 · branch covert-production
API
GET https://markdownregistry.com/api/v1/artifacts/art_vbu42jb4jaf3dqcj GET https://markdownregistry.com/api/v1/resolve?ref=anonymousnomad/aide-sovereign-workbench/aide-flagship-twin-mind-train GET https://markdownregistry.com/api/v1/blob/b03f0853e4558f0b592630c109623d6a2e3691760b6a23a3ec491ba0b05b62bf
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.