The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from the antithesis/scratchbook/ directory to guide scaffolding decisions.
Ingestion points: The skill reads antithesis/scratchbook/sut-analysis.md and antithesis/scratchbook/deployment-topology.md (specified in SKILL.md and references/docker-images.md) to determine image structure, instrumentation, and service composition.
Boundary markers: No explicit boundary markers or "ignore embedded instructions" warnings are used when interpolating or processing these external artifacts.
Capability inventory: The skill instructs the agent to write infrastructure files (Dockerfiles, docker-compose.yaml) and execute shell commands via snouty, docker, and podman (specified in references/submit-and-test.md).
Sanitization: The skill lacks explicit sanitization or strict schema validation for the contents of the research artifacts before they are used to generate code.
[EXTERNAL_DOWNLOADS]: The skill directs the agent to fetch tools and metadata from vendor-owned and well-known sources.
Fetches the snouty tool from the official Antithesis GitHub repository (github.com/antithesishq/snouty).
Queries various official package registries (NuGet, PyPI, Maven, Crates.io, and the Go Proxy) to identify the latest SDK versions for different languages.
References official Antithesis documentation pages for configuration and instrumentation guidance.
[COMMAND_EXECUTION]: The skill requires the execution of system commands to build and validate the testing harness.
Instructs the agent to run snouty doctor, snouty validate, and snouty launch (when antithesis-launch is unavailable).
Directs the use of container engine commands such as docker compose build, podman compose build, and image inspect to manage local images.
Uses the assets/antithesis/setup-complete.sh script to write setup status signals to the file system.

antithesis-setup