The Agent Skills Directory

[PROMPT_INJECTION]: The skill documents interfaces for retrieving arbitrary, user-generated content from the Arweave decentralized network, which introduces a surface for indirect prompt injection.
Ingestion points: Untrusted data enters the agent context through transaction data fields and the /tx/{id}/data.html endpoint as described in references/core-transactions.md and references/features-http-api.md.
Boundary markers: Absent; the documentation does not suggest using delimiters or specific instructions to isolate retrieved network data from the agent's internal logic.
Capability inventory: The skill enables the agent to construct and broadcast financial transactions (quantity, target) and interact with internal node APIs for signing as documented in references/features-internal-api.md.
Sanitization: Absent; there are no instructions provided for validating, filtering, or escaping content fetched from the network before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The documentation in references/features-ipfs-pinning.md references downloading the IPFS daemon from its official distribution service at dist.ipfs.io.

arweave