The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data by auditing local codebase files for specific Figma API patterns, which represents an indirect prompt injection surface.
Ingestion points: The skill performs file reading operations (via grep and Read tools) on user-controlled source files in directories like src/ to identify API endpoints and scopes.
Boundary markers: The instructions do not define clear delimiters or use "ignore embedded instructions" warnings when the agent processes the audited file content.
Capability inventory: The skill has access to sensitive capabilities including file writing (Write, Edit), network operations (curl, fetch), and shell execution (Bash).
Sanitization: No explicit sanitization or validation of the content extracted from local files is performed before it is processed by the agent.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to audit the codebase and interact with APIs.
Evidence: Extensive use of grep for pattern matching in SKILL.md and references/scope-migration-files-read-deprecation.md. It also provides curl commands for webhook registration.
[EXTERNAL_DOWNLOADS]: The skill makes network requests to external resources, specifically the Figma REST API.
Evidence: Use of fetch and curl to interact with https://api.figma.com for tasks such as creating webhooks, listing existing webhooks, and checking token health. These requests target a well-known service domain.

figma-upgrade-migration