The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from various repository files such as CONTRIBUTING.md and CHANGELOG.md to extract facts for documentation. This creates an attack surface where malicious instructions hidden in these files could influence the agent's behavior.
Ingestion points: Content is read from package.json, CONTRIBUTING.md, CHANGELOG.md, .ds-ops-config.yml, and local documentation sources using Read and Bash tools.
Boundary markers: The skill does not implement delimiters or safety instructions (e.g., 'ignore embedded commands') when interpolating the content of these files into its prompts.
Capability inventory: The agent has permissions to read files using restricted Bash commands (cat, head, etc.) and Write permissions to create the generated onboarding guide.
Sanitization: No explicit sanitization, filtering, or validation is performed on the data extracted from repository files before it is processed by the agent.

onboarding