The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a vendor-specific CLI tool to perform operations such as research collection, schema retrieval, and health checks. Specific commands include postplus research collect, postplus research schema, and postplus doctor.
[DATA_EXFILTRATION]: The skill includes functionality to retrieve public contact signals via the instagram-email-search collection key. This allows for the discovery of email addresses associated with creator profiles as part of the outreach workflow.
[PROMPT_INJECTION]: The skill creates an indirect prompt injection surface by ingesting data from external Instagram sources.
Ingestion points: Untrusted data enters the context via comments, posts, profiles, and search collection results, as seen in references/audience-voice.md and references/creator-discovery.md.
Boundary markers: The references/shared-contract.md documentation provides rules for limiting the agent to public data and instructions to separate objective facts from inferred interpretations.
Capability inventory: The agent executes shell commands using the postplus CLI (referenced in SKILL.md).
Sanitization: The skill uses structured JSON for request parameters to the CLI, but there is no explicit instruction for the sanitization of the content of the retrieved Instagram data before analysis.

instagram-research