The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use CLI tools such as manus-mcp-cli and supabase for database queries and function deployments. It specifically recommends executing SQL commands to bypass Row Level Security (RLS) for the service role when troubleshooting permission issues in references/common-deployment-issues.md. This involves using high-privilege commands to override standard security constraints.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core functionality of reading and analyzing data from external, potentially untrusted sources.
Ingestion points: The agent is directed to read database content via execute_sql and runtime logs via get_logs as specified in references/verification-checklist.md.
Boundary markers: The instructions do not define explicit delimiters or "ignore embedded instructions" directives to prevent the agent from potentially obeying instructions found within the logs or database records.
Capability inventory: The agent has the capability to execute SQL, deploy code, and perform network requests via curl.
Sanitization: There is no mention of sanitizing or validating data retrieved from the database or logs before the agent processes it.
[EXTERNAL_DOWNLOADS]: The skill involves deploying code to and interacting with Supabase. These operations target a well-known service and are documented as part of the intended verification and deployment workflow.
[SAFE]: The skill follows secure practices for credential management by instructing the use of environment variables and platform-specific secrets (e.g., OPENAI_API_KEY, SUPABASE_SERVICE_ROLE_KEY) instead of hardcoding sensitive information.

feature-verification