The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external manifest files (.jsonl) which include transcriptions and metadata. This creates an attack surface where maliciously crafted data could attempt to influence the agent's behavior.
Ingestion points: manifest.jsonl files loaded in Step 4b (split) and Step 4d (training).
Boundary markers: The instructions do not specify explicit delimiters or warnings for the agent to ignore instructions embedded in the manifest data.
Capability inventory: The skill utilizes Bash, Read, and Write tools to manage cloud instances, manipulate files, and execute Docker containers.
Sanitization: No sanitization or validation of the manifest's text content is described in the provided scripts or instructions.
[EXTERNAL_DOWNLOADS]: The skill provides instructions to download the Brev CLI installer from raw.githubusercontent.com/brevdev/brev-cli/main/bin/install-latest.sh and pulls the official NVIDIA NeMo container from nvcr.io/nvidia/nemo:25.11.01.
[COMMAND_EXECUTION]: The skill executes shell commands to provision GPU hosts via the brev CLI, manage SSH configurations, and run model training using docker run.
[CREDENTIALS_UNSAFE]: The skill notes that an NVIDIA_API_KEY is required for certain operations. This is treated as a configuration requirement for the user and no hardcoded secrets were found.

digital-health-clinical-asr-finetune