The Agent Skills Directory

[PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it is designed to ingest and summarize reports originating from external third-party integrations (e.g., GitHub, Linear, Zendesk).
Ingestion points: External data enters the agent's context through the inbox-reports-list and inbox-reports-retrieve tools referenced in SKILL.md.
Boundary markers: The instructions do not define specific delimiters or warnings to ignore instructions embedded within the processed report content.
Capability inventory: The skill utilizes read-only retrieval tools; however, the summarized data influences subsequent agent actions and user visibility.
Sanitization: There is no requirement in the workflow to sanitize or validate the content of the signal reports before the agent processes them.
[SAFE]: The skill follows secure credential management practices, explicitly instructing the agent that integration credentials are not contained within the configuration blobs and are safe to summarize.
[SAFE]: All identified tools and external links (e.g., to posthog.com) are legitimate resources owned by the author (PostHog) and align with the skill's stated purpose.

inbox-exploration