truefoundry-ml-repos（truefoundry ml 存储库）

Routing note: For ambiguous user intents, use the shared clarification templates in references/intent-clarification.md.

ML Repos

Browse TrueFoundry ML repositories and model registry. List ML repos, get repo details, and list models/artifacts within a repo.

Scope

Browse ML repositories, list models and artifacts, and retrieve FQNs for use with other skills (prompts, llm-deploy).

Step 1: Preflight

Run the status skill first to verify TFY_BASE_URL and TFY_API_KEY are set and valid.

When using direct API, set TFY_API_SH to the full path of this skill's scripts/tfy-api.sh. See references/tfy-api-setup.md for paths per agent.

Step 2: List ML Repos

Via Tool Call

tfy_ml_repos_list()

Via Direct API

TFY_API_SH=~/.claude/skills/truefoundry-ml-repos/scripts/tfy-api.sh

# List all ML repos
$TFY_API_SH GET /api/ml/v1/ml-repos

Present results:

ML Repos:
| Name          | ID       | FQN                    |
|---------------|----------|------------------------|
| my-models     | mlr-abc  | ml-repo:my-models      |
| experiment-1  | mlr-def  | ml-repo:experiment-1   |

Step 3: Get ML Repo Details

Via Tool Call

tfy_ml_repos_get(id="REPO_ID")

Via Direct API

# Get ML repo by ID
$TFY_API_SH GET /api/ml/v1/ml-repos/REPO_ID

Step 4: List Models in a Repo

Via Tool Call

tfy_models_list(ml_repo_id="REPO_ID")

Via Direct API

# List models (filter by ml_repo_id, name, or fqn)
$TFY_API_SH GET "/api/ml/v1/models?ml_repo_id=REPO_ID"

# Search by name
$TFY_API_SH GET "/api/ml/v1/models?name=my-model"

# Search by FQN
$TFY_API_SH GET "/api/ml/v1/models?fqn=model:my-models:my-model"

Present results:

Models in "my-models":
| Name        | ID       | FQN                           | Versions |
|-------------|----------|-------------------------------|----------|
| my-model    | mdl-abc  | model:my-models:my-model      | 3        |
| classifier  | mdl-def  | model:my-models:classifier    | 1        |

<success_criteria>

Success Criteria

• The user can list all ML repos and see them in a formatted table
• The user can get details for a specific ML repo by ID
• The user can list models within a repo, filtered by repo ID, name, or FQN
• The agent has provided FQN values that can be used with other skills (prompts, llm-deploy)

</success_criteria>

Composability

• Preflight: Use status skill to verify TFY_BASE_URL and TFY_API_KEY
• Prompts: ML repo FQN is needed when creating prompts (prompts skill)
• Fine-tuning: Fine-tuned model outputs are saved to ML repos
• Deploy: Models from the registry can be deployed using llm-deploy skill

API Endpoints

See references/api-endpoints.md for the full ML Repos and Models API reference.

Error Handling

ML Repo Not Found

ML repo ID not found. List repos first to find the correct ID.

No Models in Repo

This ML repo has no models yet. Models appear after logging artifacts via the SDK or fine-tuning.

Permission Denied

Cannot access ML repos. Check your API key permissions.

Empty Response

No ML repos found. Create one via the TrueFoundry UI or SDK:
  import truefoundry as tfy
  client = tfy.TrueFoundryClient()
  client.create_ml_repo(name="my-repo")