The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions do not contain any malicious patterns or attempts to bypass safety guardrails. It uses natural instructional language consistent with its purpose.
[EXTERNAL_DOWNLOADS]: The skill references official SDKs and tools from the vendor's domains. These are verified vendor resources necessary for the migration process.
[COMMAND_EXECUTION]: The skill utilizes standard utility commands such as grep and find to identify authentication touchpoints in the user's codebase. It also recommends standard package management and build commands like npm install and go build to verify the migration, which are typical for developer tools.
[CREDENTIALS_UNSAFE]: The documentation mentions environment variables such as STYTCH_SECRET and DESCOPE_MANAGEMENT_KEY in the context of configuration templates. No actual secrets are hardcoded, and the instructions emphasize secure secret management via environment variables.
[PROMPT_INJECTION]: The skill has a data ingestion surface as it reads local codebase files to generate migration plans.
Ingestion points: Codebase source files analyzed via grep and find commands.
Boundary markers: Absent from the instructional logic.
Capability inventory: Execution of shell commands (grep, find, npm, go), writing of local documentation files, and network-based testing (curl).
Sanitization: Absent; the agent relies on the user to review the generated migration plan.

stytch-to-descope