The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the repository, including source code, Software Requirements Specifications (SRS), and product documentation. This creates a surface for indirect prompt injection, as malicious instructions could be embedded in these documents to influence the agent's behavior during code modification or test generation.
Ingestion points: SRS documents, API contracts, architectural decision records, and project source files (SKILL.md, references/improve.md, references/write.md).
Boundary markers: The skill does not define specific delimiters or warnings to ignore embedded instructions in ingested documentation.
Capability inventory: The skill can write test files, modify source code to fix defects, and execute verification commands such as linting and type checking.
Sanitization: No explicit sanitization or filtering of external content is described.
[COMMAND_EXECUTION]: The skill executes various CLI tools and verification commands.
Evidence: The instructions direct the agent to run Wallaby-specific tools like analyze and inspect, as well as user-specified verification commands like lint or tsc (TypeScript compiler) to validate changes.
[DYNAMIC_EXECUTION]: The skill generates and executes code as part of its core testing workflow.
Evidence: The skill involves writing new test scripts and performing 'manual mutation testing,' which requires making temporary modifications to the source code and observing the execution results to verify test sensitivity.

wallaby