The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on executing various shell commands to manage the deployment lifecycle.
Evidence: Uses pipecat, docker, and uv commands to verify prerequisites, authenticate, manage secrets, and perform builds.
Context: These operations are central to the skill's primary purpose of deploying agents to a cloud environment.
[CREDENTIALS_UNSAFE]: The skill interacts with sensitive environment files to manage application secrets.
Evidence: Accesses .env files to pass their content to the pipecat cloud secrets set command.
Context: While this involves sensitive file access, it is a documented part of the secret management workflow and requires user-provided paths.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted configuration files from the local workspace, creating a potential injection surface.
Evidence:
Ingestion points: Reads and parses pcc-deploy.toml, Dockerfile, and uv.lock files.
Boundary markers: None identified; the skill reads these files and displays their content directly to the user/context.
Capability inventory: The skill possesses shell execution capabilities and file system access.
Sanitization: No specific validation or escaping logic for the ingested configuration data is described.
[EXTERNAL_DOWNLOADS]: The skill guides the user to install external dependencies required for operation.
Evidence: Recommends installing the pipecat-ai-cli package using uv tool install.
Context: uv is a well-known development tool, and the CLI package is the official tool for the target platform.

pipecat-deploy