The Agent Skills Directory

[PROMPT_INJECTION] (HIGH): The skill is highly susceptible to indirect prompt injection.
Ingestion points: It reads untrusted content from external Confluence pages via confluence read and confluence search commands.
Boundary markers: There are no boundary markers or instructions to isolate or ignore embedded commands within the processed data.
Capability inventory: The skill possesses powerful write capabilities, including confluence create and confluence update, as well as confluence export for data extraction.
Sanitization: No sanitization or validation of the ingested external content is performed.
[EXTERNAL_DOWNLOADS] (MEDIUM): The skill instructs the user to install a third-party package confluence-cli from npm. The repository pchuri/confluence-cli is not within the trusted source scope, making the dependency unverifiable.
[DATA_EXFILTRATION] (MEDIUM): The skill provides capabilities to export entire Confluence pages and their attachments to the local filesystem (confluence export), which could be used to facilitate unauthorized data removal if the agent is compromised by an injection.

confluence