The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon external data from Notion's developer documentation to guide code migrations, which presents a potential indirect prompt injection surface.
Ingestion points: The skill uses Read and Grep to analyze documentation referenced in references/official-docs.md (linking to developers.notion.com) and local repository code as part of the inventory process (SKILL.md instructions, Step 1 and 2).
Boundary markers: The skill contains explicit safety instructions in the 'Tool Discipline' section of SKILL.md and the 'Evidence rules' section of references/official-docs.md, stating that invocation does not grant authorization for network access, credentials, or deployment.
Capability inventory: The skill is granted Write and Edit tools to modify repository artifacts based on the analysis of the documentation.
Sanitization: The instructions mandate the use of "synthetic or approved non-production fixtures with secrets and workspace content removed" and require that the agent "preserve redacted request evidence" if discrepancies are found.

notion-upgrade-migration