The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes code and documentation from the target repository, creating a surface for indirect instructions. 1. Ingestion points: Git diff content, documentation in docs/fixes/, and PRD files. 2. Boundary markers: No explicit markers are used to separate ingested data from agent instructions. 3. Capability inventory: Executes subprocesses, uses GitHub CLI, and writes to the local .context/ directory. 4. Sanitization: No sanitization is applied to the repository content before analysis.
[COMMAND_EXECUTION]: The skill invokes several command-line tools to manage the testing lifecycle. Evidence: Uses gh api for version resolution, git for diffing, and go tool mage for execution.
[DYNAMIC_EXECUTION]: The skill compiles Atmos binaries from source code at runtime. Evidence: Builds specific commit SHAs into binaries to perform behavior comparisons in a controlled environment.

regression-test