The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to perform repository management and mutation testing using local shell commands. Evidence: File 'references/mutations.md' specifies the use of 'git worktree add --detach', 'git worktree remove --force ', and 'git worktree prune' to create and clean up temporary test environments.
[INDIRECT_PROMPT_INJECTION]: The skill consumes untrusted data from external sources while possessing the capability to execute commands and modify the codebase. Ingestion points: In 'SKILL.md', the agent is instructed to read 'the spec, or the request when there is none, the diff range, and the assertion changes' as primary inputs for the review process. Boundary markers: The instructions lack specific directives for the agent to use secure delimiters or to disregard natural language instructions that might be embedded within these external inputs. Capability inventory: The agent is tasked with modifying code to fix identified issues and executing shell commands for git operations. Sanitization: There is no mention of input validation or sanitization routines for the data processed during the review phase.

refine