The Agent Skills Directory

[SAFE]: No malicious code, obfuscation, hardcoded credentials, or unauthorized network operations were detected in the skill files. The instructions promote architectural best practices and standard developer workflows.
[COMMAND_EXECUTION]: The skill instructs the agent to execute standard development tools such as tsc (TypeScript compiler), linters, and formatters to verify fixes applied during the audit process. These commands are executed locally within the project context.
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection because it is designed to ingest and analyze untrusted source code from routes, feature slices, and pull requests. While acting on this data, the agent has capabilities that could be abused if the code contains malicious instructions.
Ingestion points: Source code from routes, feature slices, PRs, and components as defined in SKILL.md.
Boundary markers: Absent; there are no delimiters or instructions to ignore embedded commands within the analyzed code.
Capability inventory: File read/write access and shell command execution (tsc, lint, formatter) as described in SKILL.md.
Sanitization: Absent; the skill does not specify validation or sanitization of the audited code for malicious instructions.

audit-implementation