@tangle-network/sandbox-cli 0.22.79 → 0.22.80

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (3) hide show
  1. package/README.md +8 -8
  2. package/SKILL.md +10 -9
  3. package/package.json +6 -6
package/README.md CHANGED
@@ -248,8 +248,8 @@ tangle agent prompt sbx_123 "file the report" --connection github:* --allow-writ
248
248
 
249
249
  # temporary GPU for an eval; omitted provider picks the cheapest configured cloud
250
250
  tangle sandbox gpu run sbx_123 \
251
- --accelerator-kind nvidia-3090 \
252
- --accelerator-memory 24000 \
251
+ --accelerator-kind nvidia-rtx-3090 \
252
+ --accelerator-memory 24 \
253
253
  --max-spend-usd 1 \
254
254
  --max-lifetime 900 \
255
255
  --idle-timeout 120 \
@@ -276,8 +276,8 @@ Always set a spend cap and lifetime.
276
276
  ```bash
277
277
  # Attach, run one command, and destroy the lease.
278
278
  tangle sandbox gpu run sbx_123 \
279
- --accelerator-kind nvidia-3090 \
280
- --accelerator-memory 24000 \
279
+ --accelerator-kind nvidia-rtx-3090 \
280
+ --accelerator-memory 24 \
281
281
  --max-spend-usd 1 \
282
282
  --max-lifetime 900 \
283
283
  --idle-timeout 120 \
@@ -288,8 +288,8 @@ For multiple GPU commands, use the manual lifecycle:
288
288
 
289
289
  ```bash
290
290
  tangle sandbox gpu attach sbx_123 \
291
- --accelerator-kind nvidia-3090 \
292
- --accelerator-memory 24000 \
291
+ --accelerator-kind nvidia-rtx-3090 \
292
+ --accelerator-memory 24 \
293
293
  --max-spend-usd 1 \
294
294
  --max-lifetime 900 \
295
295
  --idle-timeout 120
@@ -303,8 +303,8 @@ Create-time GPU flags are a shortcut over the same lease lifecycle:
303
303
 
304
304
  ```bash
305
305
  tangle sandbox create --name gpu-eval \
306
- --accelerator-kind nvidia-3090 \
307
- --accelerator-memory 24000 \
306
+ --accelerator-kind nvidia-rtx-3090 \
307
+ --accelerator-memory 24 \
308
308
  --gpu-max-spend-usd 1 \
309
309
  --gpu-max-lifetime 900 \
310
310
  --gpu-idle-timeout 120
package/SKILL.md CHANGED
@@ -126,17 +126,18 @@ tangle sandbox gpu providers <sandbox-id> --accelerator-kind nvidia-h100
126
126
  | AWS | `aws` | `AWS_ACCESS_KEY_ID`, `AWS_SECRET_ACCESS_KEY`, `AWS_REGION` |
127
127
  | GCP | `gcp` | `GCP_PROJECT_ID`, `GCP_SERVICE_ACCOUNT_KEY` |
128
128
  | Azure | `azure` | `AZURE_TENANT_ID`, `AZURE_CLIENT_ID`, `AZURE_CLIENT_SECRET`, `AZURE_SUBSCRIPTION_ID`, `AZURE_RESOURCE_GROUP`, `AZURE_SUBNET_ID` |
129
- | Modal | `modal` | `MODAL_TOKEN_ID`, `MODAL_TOKEN_SECRET` |
129
+ | Crusoe | `crusoe` | `CRUSOE_ACCESS_KEY_ID`, `CRUSOE_SECRET_KEY`, `CRUSOE_PROJECT_ID`, `CRUSOE_GPU_HOURLY_USD` |
130
+ | Bittensor Lium | `bittensor-lium` | `LIUM_API_KEY` |
130
131
 
131
- Setting a provider's variables enables it; `GPU_LEASE_PROVIDERS` only narrows the set.
132
+ `GPU_LEASE_PROVIDERS` lists the providers an orchestrator offers (production lists all of them); a listed provider is enabled by its variables.
132
133
  An attach with no usable provider fails with 503 `GPU_PROVIDER_NOT_CONFIGURED` naming the variables to set; a refused key shows as `credential-invalid` and is never selected.
133
- Modal offers T4, L4, A10G, L40S, A100, H100, H200 and B200, not consumer cards such as the RTX 3090.
134
+ Accelerator kinds are the canonical GPU class ids such as `nvidia-rtx-3090`, `nvidia-a100-80gb` and `nvidia-h100`; `--accelerator-memory` is decimal GB per GPU.
134
135
 
135
136
  ```bash
136
137
  # One-shot eval: attach GPU, run command, always destroy the lease.
137
138
  tangle sandbox gpu run <sandbox-id> \
138
- --accelerator-kind nvidia-3090 \
139
- --accelerator-memory 24000 \
139
+ --accelerator-kind nvidia-rtx-3090 \
140
+ --accelerator-memory 24 \
140
141
  --max-spend-usd 1 \
141
142
  --max-lifetime 900 \
142
143
  --idle-timeout 120 \
@@ -144,8 +145,8 @@ tangle sandbox gpu run <sandbox-id> \
144
145
 
145
146
  # Manual lifecycle when the agent needs multiple GPU commands.
146
147
  tangle sandbox gpu attach <sandbox-id> \
147
- --accelerator-kind nvidia-3090 \
148
- --accelerator-memory 24000 \
148
+ --accelerator-kind nvidia-rtx-3090 \
149
+ --accelerator-memory 24 \
149
150
  --max-spend-usd 1 \
150
151
  --max-lifetime 900 \
151
152
  --idle-timeout 120
@@ -157,8 +158,8 @@ For fleets, put the same GPU lease cap on every worker:
157
158
 
158
159
  ```bash
159
160
  tangle fleet create --count 4 \
160
- --accelerator-kind nvidia-3090 \
161
- --accelerator-memory 24000 \
161
+ --accelerator-kind nvidia-rtx-3090 \
162
+ --accelerator-memory 24 \
162
163
  --gpu-max-spend-usd 1 \
163
164
  --gpu-max-lifetime 900 \
164
165
  --gpu-idle-timeout 120
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@tangle-network/sandbox-cli",
3
- "version": "0.22.79",
3
+ "version": "0.22.80",
4
4
  "description": "CLI for Tangle Sandbox operations",
5
5
  "type": "module",
6
6
  "engines": {
@@ -26,7 +26,7 @@
26
26
  "@tangle-network/agent-runtime": "^9.0.3",
27
27
  "@tangle-network/harness-sessions": "^1.0.2",
28
28
  "@tangle-network/hub-sdk": "1.0.26",
29
- "@tangle-network/sandbox": "^1.0.357",
29
+ "@tangle-network/sandbox": "^1.0.358",
30
30
  "@tangle-network/sdk-cli-runner": "^0.3.12",
31
31
  "chalk": "^5.6.2",
32
32
  "cli-bridge": "https://github.com/drewstone/cli-bridge/archive/6c215f6912dc65ab6145d4102fee744d7fd09b79.tar.gz",
@@ -59,8 +59,8 @@
59
59
  "tsx": "^4.23.15",
60
60
  "typescript": "^6.0.3",
61
61
  "vitest": "^4.1.12",
62
- "@repo/shared": "0.0.0",
63
- "@tangle-network/runtime-contracts": "0.20.0"
62
+ "@tangle-network/runtime-contracts": "0.20.0",
63
+ "@repo/shared": "0.0.0"
64
64
  },
65
65
  "publishConfig": {
66
66
  "access": "public",
@@ -84,8 +84,8 @@
84
84
  "author": "Tangle Network",
85
85
  "license": "MIT",
86
86
  "tangleRelease": {
87
- "sourceSha": "ffcb0cb89f57dc90806f698e4681a4d44008d9d7",
88
- "mergedAt": "2026-11-09T23:32:43-08:00"
87
+ "sourceSha": "a8b4845ca412f331d1089273862d166ae8f60c95",
88
+ "mergedAt": "2026-11-10T11:02:57-08:00"
89
89
  },
90
90
  "scripts": {
91
91
  "build": "tsdown",