The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands like grep to perform comprehensive scans of the local filesystem to identify authentication logic, environment variables, and dependency manifests. It also executes npm view to verify package versions during the analysis phase.
[EXTERNAL_DOWNLOADS]: The skill recommends that the agent or user install the 'Descope Docs MCP' from a vendor-controlled domain (docs-mcp.descope.com). This is categorized as a safe vendor resource but involves a remote tool dependency.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting and processing untrusted data from the user's project files to generate migration plans.
Ingestion points: Codebase analysis results, including file contents and grep matches, are read into the agent's context (SKILL.md).
Boundary markers: The skill requires the agent to generate a MIGRATION-PLAN.md for human review and uses AskUserQuestion to confirm architecture decisions, providing a manual oversight step.
Capability inventory: The skill possesses the capability to execute shell commands (grep, npm view) and write multiple files to the local directory (MIGRATION-PLAN.md, MIGRATION-STATE.md, MIGRATION-SUMMARY.md).
Sanitization: The instructions do not specify automated sanitization or filtering of the ingested codebase content before it is processed by the LLM.

stytch-to-descope