prose-contracts

221 Commits 134 Branches 0 Tags

3 Commits

Include renames

Author	SHA1	Message	Date
root	d646539e81	FLEET-WIDE STALE gemma-4-12b PURGE — replace with gpu-vision Purge all gemma-4-12b references from prose-contracts repo. Current reality (verified 2026-09-06): - .8 (ct8, RTX 3090): Qwen3.8-27B-Uncensored-Q4_K_M, aliases: qwen3.6-27B-code, gpu-dense - .15 (strix): Carnice-Qwen3.6-35B-A3B-Q4_K_M MoE, alias: strix-moe - .110 (ct110, RTX 5070): Qwen3.5-9B VISION (multimodal), alias: gpu-vision Syslog-auto weighted pool: qwen3.6-27B-code 0.70 + strix-moe 0.20 + gpu-vision 0.10 All gemma-4-12b references replaced with gpu-vision where it was the RTX 5070 model. Vision/light roles now route to gpu-vision (not gemma-4-12b).	2026-09-06 12:19:27 +00:00
root	cd479caeec	Fix: update compression model to syslog-auto across contract and audit (Rule 7) - Updated hermes-config-template.prose.md: all references to strix-moe for compression changed to syslog-auto to match operational decision on 2026-07-23 (prevents sustained Strix Halo thermal load via weighted pool). - Updated audit-hermes-config.py Rule 7 to expect syslog-auto instead of strix-moe, ensuring Abiba's next run validates against the correct baseline.	2026-07-23 18:02:56 +00:00
root	1b1de8b0fc	Add Rule 14 (provider name must match custom_providers) + audit script Rule 14: model.provider MUST be 'harness' (custom_providers[0].name), NOT 'custom'. When provider: custom, Hermes falls through to generic resolution path that ignores key_env, producing 'no-key-required' → HTTP 401. audit-hermes-config.py: encodes all 14 contract rules as automated checks. Run before and after any Hermes config change. Root cause: WAL #1471 (2026-07-19 Mumuni 401 incident)	2026-07-23 12:16:33 +00:00

Author

SHA1

Message

Date

root

d646539e81

FLEET-WIDE STALE gemma-4-12b PURGE — replace with gpu-vision

Purge all gemma-4-12b references from prose-contracts repo.

Current reality (verified 2026-09-06):
- .8 (ct8, RTX 3090): Qwen3.8-27B-Uncensored-Q4_K_M, aliases: qwen3.6-27B-code, gpu-dense
- .15 (strix): Carnice-Qwen3.6-35B-A3B-Q4_K_M MoE, alias: strix-moe
- .110 (ct110, RTX 5070): Qwen3.5-9B VISION (multimodal), alias: gpu-vision

Syslog-auto weighted pool: qwen3.6-27B-code 0.70 + strix-moe 0.20 + gpu-vision 0.10

All gemma-4-12b references replaced with gpu-vision where it was the RTX 5070 model.
Vision/light roles now route to gpu-vision (not gemma-4-12b).

2026-09-06 12:19:27 +00:00

root

cd479caeec

Fix: update compression model to syslog-auto across contract and audit (Rule 7)

- Updated hermes-config-template.prose.md: all references to strix-moe for
  compression changed to syslog-auto to match operational decision on 2026-07-23
  (prevents sustained Strix Halo thermal load via weighted pool).
- Updated audit-hermes-config.py Rule 7 to expect syslog-auto instead of
  strix-moe, ensuring Abiba's next run validates against the correct baseline.

2026-07-23 18:02:56 +00:00

root

1b1de8b0fc

Add Rule 14 (provider name must match custom_providers) + audit script

Rule 14: model.provider MUST be 'harness' (custom_providers[0].name), NOT 'custom'.
When provider: custom, Hermes falls through to generic resolution path that
ignores key_env, producing 'no-key-required' → HTTP 401.

audit-hermes-config.py: encodes all 14 contract rules as automated checks.
Run before and after any Hermes config change.

Root cause: WAL #1471 (2026-07-19 Mumuni 401 incident)

2026-07-23 12:16:33 +00:00