The Agent Skills Directory

[SAFE]: The skill includes a validation checklist that encourages scanning the source code for hardcoded secrets and connection strings using grep, which aligns with security best practices.
[SAFE]: Secret management is handled correctly by advising the use of environment variables and .env files, while specifically warning against hardcoding credentials.
[SAFE]: Network activity is restricted to local endpoint testing via curl, and build processes utilize standard, well-recognized package managers and test runners.
[INDIRECT_PROMPT_INJECTION]: The agent is instructed to ingest and process external data sources such as API contracts and shared types, which represents a potential attack surface.
Ingestion points: Data is read from 'contracts/' and 'shared/' directories.
Boundary markers: The skill does not define explicit delimiters or instructions to ignore embedded commands within these files.
Capability inventory: The agent has access to Bash and file writing tools, which could be exploited if malicious content were processed.
Sanitization: No specific sanitization or validation logic is prescribed for the content of the contracts beyond implementation matching.

backend-agent