The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes git commands via execFileSync in assets/scripts/mutation-plan.mjs to identify changed files and hunks for mutation testing.
Evidence: The git helper function uses the array form of execFileSync to prevent shell injection and employs the --end-of-options flag when resolving the comparison base to safeguard against argument injection.
[INDIRECT_PROMPT_INJECTION]: The skill processes Git diffs and mutation reports (reports/mutation/mutation.json) to generate summaries for AI review, creating a surface where malicious content in source code or test names could influence agent behavior.
Ingestion points: assets/scripts/mutation-summary.mjs reads and parses Stryker JSON reports; assets/scripts/mutation-plan.mjs reads output from Git diff commands.
Boundary markers: The skill does not currently employ explicit delimiters or "ignore embedded instructions" warnings when presenting extracted code snippets or test names to the agent.
Capability inventory: The skill allows for file system access, system command execution (git, pnpm), and network operations via the defined CI workflow.
Sanitization: Extracted data from the reports (such as source code fragments and test identifiers) is parsed but not sanitized or escaped before inclusion in the summary for the AI.
[DYNAMIC_EXECUTION]: The script assets/scripts/mutation-changed.mjs performs a dynamic import() of the @stryker-mutator/core library at runtime to execute mutation tests.

stryker-js