The Agent Skills Directory

[PROMPT_INJECTION]: The skill demonstrates workflows that ingest data from the Notion REST API, creating a potential surface for indirect prompt injection. If a malicious actor can edit shared Notion content, they could attempt to influence the agent's subsequent actions. * Ingestion points: Notion API responses from database queries and page retrievals as documented in references/notion-api-patterns.md and examples/product-team-workspace-buildout.md. * Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are included in the provided templates. * Capability inventory: Capabilities are limited to Notion API interactions (creating/updating pages); the skill does not include subprocess calls or local file-system writes. * Sanitization: No explicit sanitization or validation of external Notion content is described in the prompt instructions.
[EXTERNAL_DOWNLOADS]: The skill correctly references and integrates with well-known professional services including Notion, Jira, Linear, and GitHub using official API documentation and standard integration endpoints.
[SAFE]: The skill follows security best practices by using clearly marked placeholders (secret_...) for authentication tokens and includes a dedicated 'Red Flags' guide (references/red-flags.md) that explicitly warns against storing sensitive credentials within Notion pages.

notion-pm