The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from repository files including delivery-plan.md, git diff output, and local application responses via curl. These sources could theoretically contain instructions intended to influence the agent's behavior. The skill lacks explicit boundary markers or sanitization for this external content, but it operates within a standard development workflow expected for an automated testing tool.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to perform repository analysis and environment detection. The commands are scoped to local operations such as git plumbing, filesystem checks (ls), and connectivity checks to localhost ports (curl localhost). These operations are consistent with the skill's stated purpose of identifying changes and verifying a running application.

test-and-fix