METR
Official@metr
Model Evaluation and Threat Research
Agent Skills by METR
Showing 9 vetted skills indexed across 2 GitHub repositories.
METRrepo-sync
Cherry-pick commits from legacy repositories into the hawk monorepo and port Terraform changes to Pulumi.
METRfull-stack-dev
Coordinate frontend and backend development for Hawk's web UI and API.
METRview-results
List, inspect, and retrieve detailed results from Hawk AI evaluations.
METRfullstack-dev
Manage frontend and backend services with local library linking.
METRdebug-stuck-eval
Diagnose stalled AI evaluations in the Hawk/Inspect framework.
METRdeploy-dev
Deploy code and infrastructure updates to development environments using Terraform.
METRsmoke-tests
Execute automated smoke tests against deployed environments using pytest.
METRmonitoring
Monitor Hawk job status and retrieve logs for debugging.
METRdatabase-migrations
Manage database schema evolution with Alembic for Python applications.
Frequently Asked Questions About METR
FAQPage SchemaWhat specific tasks does METR enable for evaluation engineers?▼
METR enables engineers to list, inspect, and retrieve granular results from evaluation runs. It provides specific capabilities for diagnosing stalled evaluation processes within the Hawk framework and monitoring job statuses through centralized log retrieval to ensure high-fidelity assessment cycles.
Which infrastructure management tasks are supported by these capabilities?▼
These capabilities support the deployment of code and infrastructure updates to development environments using Terraform. It includes porting legacy repository commits into monorepos and migrating existing infrastructure definitions to Pulumi for improved state management and environment consistency.
What are the prerequisites for managing database schema evolution?▼
Managing database schema evolution requires an environment configured with Alembic. Users must ensure the target application environment is correctly linked to the local library structure to facilitate schema versioning and migration execution across the backend services.